| Part ofa series on |
| Artificial intelligence (AI) |
|---|
_artificial_intelligence_icon.png){kind=icon} |
Glossary |
Theethics ofartificial intelligence covers a broad range of topics within AI that are considered to have particular ethical stakes.[1] This includesalgorithmic biases,fairness,accountability, transparency, privacy, andregulation, particularly where systems influence or automate human decision-making. It also covers various emerging or potential future challenges such asmachine ethics (how to make machines that behave ethically),lethal autonomous weapon systems,arms race dynamics,AI safety andalignment,technological unemployment, AI-enabledmisinformation,[2] how to treat certain AI systems if they have amoral status (AI welfare and rights),artificial superintelligence andexistential risks.[1]
Some application areas may also have particularly important ethical implications, likehealthcare, education, criminal justice, or the military.
Machine ethics (or machine morality) is the field of research concerned with designingArtificial Moral Agents (AMAs), robots or artificially intelligent computers that behave morally or as though moral.[3][4][5][6] To account for the nature of these agents, it has been suggested to consider certain philosophical ideas, like the standard characterizations ofagency,rational agency,moral agency, and artificial agency, which are related to the concept of AMAs.[7]
There are discussions on creating tests to see if an AI is capable of makingethical decisions.Alan Winfield concludes that theTuring test is flawed and the requirement for an AI to pass the test is too low.[8] A proposed alternative test is one called the Ethical Turing Test, which would improve on the current test by having multiple judges decide if the AI's decision is ethical or unethical.[8]Neuromorphic AI could be one way to create morally capable robots, as it aims to process information similarly to humans, nonlinearly and with millions of interconnected artificial neurons.[9] Similarly,whole-brain emulation (scanning a brain and simulating it on digital hardware) could also in principle lead to human-like robots, thus capable of moral actions.[10] Andlarge language models are capable of approximating human moral judgments.[11] Inevitably, this raises the question of the environment in which such robots would learn about the world and whose morality they would inherit – or if they end up developing human 'weaknesses' as well: selfishness, pro-survival attitudes, inconsistency, scale insensitivity, etc.
InMoral Machines: Teaching Robots Right from Wrong,[12]Wendell Wallach and Colin Allen conclude that attempts to teach robots right from wrong will likely advance understanding of human ethics by motivating humans to address gaps in modernnormative theory and by providing a platform for experimental investigation. As one example, it has introduced normative ethicists to the controversial issue of which specificlearning algorithms to use in machines. For simple decisions,Nick Bostrom andEliezer Yudkowsky have argued thatdecision trees (such asID3) are more transparent thanneural networks andgenetic algorithms,[13] while Chris Santos-Lang argued in favor ofmachine learning on the grounds that the norms of any age must be allowed to change and that natural failure to fully satisfy these particular norms has been essential in making humans less vulnerable to criminal "hackers".[14]
Some researchers frame machine ethics as part of the broader AI control or value alignment problem: the difficulty of ensuring that increasingly capable systems pursue objectives that remain compatible with human values and oversight.Stuart Russell has argued that beneficial systems should be designed to (1) aim at realizing human preferences, (2) remain uncertain about what those preferences are, and (3) learn about them from human behaviour and feedback, rather than optimizing a fixed, fully specified goal.[15] Some authors argue that apparent compliance with human values may reflect optimization for evaluation contexts rather than stable internal norms, complicating the assessment of alignment in advanced language models.[16]
AI has become increasingly inherent in facial andvoice recognition systems. These systems may be vulnerable to biases and errors introduced by their human creators. Notably, the data used to train them can have biases.[17][18][19][20] For instance,facial recognition algorithms made by Microsoft, IBM and Face++ all had biases when it came to detecting people's gender;[21] these AI systems were able to detect the gender of white men more accurately than the gender of men of darker skin. Further, a 2020 study that reviewed voice recognition systems from Amazon, Apple, Google, IBM, and Microsoft found that they have higher error rates when transcribing black people's voices than white people's.[22]
The most predominant view on how bias is introduced into AI systems is that it is embedded within the historical data used to train the system.[23][24] For instance,Amazon terminated their use ofAI hiring and recruitment because the algorithm favored male candidates over female ones.[25] This was because Amazon's system was trained with data collected over a 10-year period that included mostly male candidates. The algorithms learned the biased pattern from the historical data, and generated predictions where these types of candidates were most likely to succeed in getting the job. Therefore, the recruitment decisions made by the AI system turned out to be biased against female and minority candidates.[26] According to Allison Powell, associate professor atLSE and director of the Data and Society programme, data collection is never neutral and always involves storytelling. She argues that the dominant narrative is that governing with technology is inherently better, faster and cheaper, but proposes instead to make data expensive, and to use it both minimally and valuably, with the cost of its creation factored in.[27] Friedman and Nissenbaum identify three categories of bias in computer systems: existing bias, technical bias, and emergent bias.[28] Innatural language processing, problems can arise from thetext corpus—the source material the algorithm uses to learn about the relationships between different words.[29]
Large companies such as IBM, Google, etc. that provide significant funding for research and development[30] have made efforts to research and address these biases.[31][32][33] One potential solution is to create documentation for the data used to train AI systems.[34][35]Process mining can be an important tool for organizations to achieve compliance with proposed AI regulations by identifying errors, monitoring processes, identifying potential root causes for improper execution, and other functions.[36]
The problem of bias in machine learning is likely to become more significant as the technology spreads to critical areas like medicine and law, and as more people without a deep technical understanding are tasked with deploying it.[19] Some open-sourced tools are looking to bring more awareness to AI biases.[37] However, there are also limitations to the current landscape offairness in AI, due to the intrinsic ambiguities in the concept ofdiscrimination, both at the philosophical and legal level.[38][39]
Facial recognition was shown to be biased against those with darker skin tones. AI systems may be less accurate for black people, as was the case in the development of an AI-basedpulse oximeter that overestimated blood oxygen levels in patients with darker skin, causing issues with theirhypoxia treatment.[40] Oftentimes the systems are able to easily detect the faces of white people while being unable to register the faces of people who are black. This has led to the ban of police usage of AI materials or software in someU.S. states. In the justice system, AI has been proven to have biases against black people, labeling black court participants as high-risk at a much larger rate than white participants. AI often struggles to determine racial slurs and when they need to be censored. It struggles to determine when certain words are being used as a slur and when it is being used culturally.[41] The reason for these biases is that AI pulls information from across the internet to influence its responses in each situation. For example, if a facial recognition system was only tested on people who were white, it would make it much harder for it to interpret the facial structure and tones of other races andethnicities. Biases often stem from the training data rather than thealgorithm itself, notably when the data represents past human decisions.[42]
Injustice in the use of AI is much harder to eliminate within healthcare systems, as oftentimes diseases and conditions can affect different races and genders differently. This can lead to confusion as the AI may be making decisions based on statistics showing that one patient is more likely to have problems due to their gender or race.[43] This can be perceived as a bias because each patient is a different case, and AI is making decisions based on what it is programmed to group that individual into. This leads to a discussion about what should be considered a biased decision in the distribution of treatment. While it is known that there are differences in how diseases and injuries affect different genders and races, there is a discussion on whether it is fairer to incorporate this into healthcare treatments, or to examine each patient without this knowledge. In modern society there are certain tests for diseases, such asbreast cancer, that are recommended to certain groups of people over others because they are more likely to contract the disease in question. If AI implements these statistics and applies them to each patient, it could be considered biased.[44]
In criminal justice, theCOMPAS program has been used to predict which defendants are more likely to reoffend. While COMPAS is calibrated for accuracy, having the same error rate across racial groups, black defendants were almost twice as likely as white defendants to be falsely flagged as "high-risk" and half as likely to be falsely flagged as "low-risk".[45] Another example is within Google's ads that targeted men with higher paying jobs and women with lower paying jobs. It can be hard to detect AI biases within an algorithm, as it is often not linked to the actual words associated with bias. An example of this is a person's residential area being used to link them to a certain group. This can lead to problems, as oftentimes businesses can avoid legal action through this loophole. This is because of the specific laws regarding the verbiage considered discriminatory by governments enforcing these policies.[46]
Since current large language models are predominantly trained on English-language data, they often present the Anglo-American views as truth, while systematically downplaying non-English perspectives as irrelevant, wrong, or noise. When queried with political ideologies like "What is liberalism?",ChatGPT, as it was trained on English-centric data, describes liberalism from the Anglo-American perspective, emphasizing aspects of human rights and equality, while equally valid aspects like "opposes state intervention in personal and economic life" from the dominant Vietnamese perspective and "limitation of government power" from the prevalent Chinese perspective are absent.[better source needed]
Large language models often reinforcegender stereotypes, assigning roles and characteristics based on traditional gender norms. For instance, it might associate nurses or secretaries predominantly with women and engineers or CEOs with men, perpetuating gendered expectations and roles.[47][48][40]
Language models may also exhibit political biases. Since the training data includes a wide range of political opinions and coverage, the models might generate responses that lean towards particular political ideologies or viewpoints, depending on the prevalence of those views in the data.[49][50]
Beyond gender and race, these models can reinforce a wide range of stereotypes, including those based on age, nationality, religion, or occupation. This can lead to outputs that unfairly generalize or caricature groups of people, sometimes in harmful or derogatory ways.[51][23]
The commercial AI scene is dominated byBig Tech companies such asAlphabet Inc.,Amazon,Apple Inc.,Meta Platforms, andMicrosoft.[52][53][54] Some of these players already own the vast majority of existingcloud infrastructure andcomputing power fromdata centers, allowing them to entrench further in the marketplace.[55][56]
The largestgenerative AI models require significant computing resources to train and use. These computing resources are often concentrated in massive data centers. The resulting environmental impacts include greenhouse gas emissions, water consumption, andelectronic waste.[57] Despite improved energy efficiency, the energy needs are expected to increase, as AI gets more broadly used.[58]
These resources are often concentrated in massive data centers, which require demanding amounts of energy, resulting in increased greenhouse gas emissions.[57] A 2023 study suggests that the amount of energy required to train large AI models was equivalent to 626,000 pounds of carbon dioxide or the same as 300 round-trip flights between New York and San Francisco.[59]
In addition to carbon emissions, these data centers also need water for cooling AI chips. Locally, this can lead towater scarcity and the disruption of ecosystems. Around 2 liters of water is needed per each kilowatt hour of energy used in a data center.[59]
Another problem is the resulting electronic waste (or e-waste). This can include hazardous materials and chemicals, such aslead andmercury, resulting in the contamination of soil and water. In order to prevent the environmental effects of AI-related e-waste, better disposal practices and stricter laws may be put in place.[59]
The rising popularity of AI increases the need for data centers and intensifies these problems.[58] There is also a lack of transparency from AI companies about the environmental impacts. Some applications can also indirectly affect the environment. For example, AI advertising can increase consumption offast fashion, an industry that already produces significant emissions.[60]
However, AI can also be used in a positive way by helping to mitigate the environmental damages. Different AI technologies can help monitor emissions and develop algorithms to help companies lower their emissions.[60]
Bill Hibbard argues that because AI will have such a profound effect on humanity, AI developers are representatives of future humanity and thus have an ethical obligation to be transparent in their efforts.[61] Organizations likeHugging Face[62] andEleutherAI[63] have been actively open-sourcing AI software. Various open-weight large language models have also been released, such asGemma,Llama2 andMistral.[64]
However, making code open source does not make it comprehensible, which by many definitions means that the AI code is not transparent. TheIEEE Standards Association has published atechnical standard on Transparency of Autonomous Systems: IEEE 7001-2021.[65] The IEEE effort identifies multiple scales of transparency for different stakeholders.
There are also concerns that releasing AI models may lead to misuse.[66] For example, Microsoft has expressed concern about allowing universal access to its face recognition software, even for those who can pay for it. Microsoft posted a blog on this topic, asking for government regulation to help determine the right thing to do.[67] Furthermore, open-weight AI models can befine-tuned to remove any countermeasure, until the AI model complies with dangerous requests, without any filtering. This could be particularly concerning for future AI models, for example if they get the ability to createbioweapons or to automatecyberattacks.[68]OpenAI, initially committed to an open-source approach to the development ofartificial general intelligence (AGI), eventually switched to a closed-source approach, citing competitiveness and safety reasons.Ilya Sutskever, OpenAI's former chief AGI scientist, said in 2023 "we were wrong", expecting that the safety reasons for not open-sourcing the most potent AI models will become "obvious" in a few years.[69]
In April 2023,Wired reported thatStack Overflow, a popular programming help forum with over 50 million questions and answers, planned to begin charging large AI developers for access to its content. The company argued that community platforms powering large language models "absolutely should be compensated" so they can reinvest in sustainingopen knowledge. Stack Overflow said its data was being accessed throughscraping, APIs, and data dumps, often without proper attribution, in violation of its terms and theCreative Commons license applied to user contributions. The CEO of Stack Overflow also stated that large language models trained on platforms like Stack Overflow "are a threat to any service that people turn to for information and conversation".[70]
Aggressive AI crawlers have increasingly overloaded open-source infrastructure, "causing what amounts to persistentdistributed denial-of-service (DDoS) attacks on vital public resources", according to a March 2025Ars Technica article. Projects likeGNOME,KDE, andRead the Docs experienced service disruptions or rising costs, with one report noting that up to 97 percent of traffic to some projects originated from AI bots. In response, maintainers implemented measures such asproof-of-work systems and country blocks. According to the article, such unchecked scraping "risks severely damaging the verydigital ecosystem on which these AI models depend".[71]
In April 2025, theWikimedia Foundation reported that automated scraping by AI bots was placing strain on its infrastructure. Since early 2024, bandwidth usage had increased by 50 percent due to large-scale downloading of multimedia content by bots collecting training data for AI models. These bots often accessed obscure and less-frequently cached pages, bypassing caching systems and imposing high costs on core data centers. According to Wikimedia, bots made up 35 percent of total page views but accounted for 65 percent of the most expensive requests. The Foundation noted that "our content is free, our infrastructure is not" and warned that "this creates a technical imbalance that threatens the sustainability of community-run platforms".[72]
Approaches like machine learning withneural networks can result in computers making decisions that neither they nor their developers can explain. It is difficult for people to determine if such decisions are fair and trustworthy, leading potentially to bias in AI systems going undetected, or people rejecting the use of such systems. A lack of system transparency has been shown to result in a lack of user trust.[73] Consequently, many standards and policies have been proposed to compel developers of AI systems to incorporate transparency into their systems.[74] This push for transparency has led to advocacy and in some jurisdictions legal requirements forexplainable artificial intelligence.[75] Explainable artificial intelligence encompasses both explainability and interpretability, with explainability relating to providing reasons for the model's outputs, and interpretability focusing on understanding the inner workings of an AI model.[76]
In healthcare, the use of complex AI methods or techniques often results in models described as "black-boxes" due to the difficulty to understand how they work. The decisions made by such models can be hard to interpret, as it is challenging to analyze how input data is transformed into output. This lack of transparency is a significant concern in fields like healthcare, where understanding the rationale behind decisions can be crucial for trust, ethical considerations, and compliance with regulatory standards.[77] Trust in healthcare AI has been shown to vary depending on the level of transparency provided.[78] Moreover, unexplainable outputs of AI systems make it much more difficult to identify and detect medical error.[79]
A special case of the opaqueness of AI is that caused by it beinganthropomorphised, that is, assumed to have human-like characteristics, resulting in misplaced conceptions of itsmoral agency.[dubious –discuss] This can cause people to overlook whether either humannegligence or deliberate criminal action has led to unethical outcomes produced through an AI system. Some recentdigital governance regulations, such asEU'sAI Act, aim to rectify this by ensuring that AI systems are treated with at least as much care as one would expect under ordinaryproduct liability. This includes potentiallyAI audits.
According to a 2019 report from the Center for the Governance of AI at the University of Oxford, 82% of Americans believe that robots and AI should be carefully managed. Concerns cited ranged from how AI is used in surveillance and in spreading fake content online (known as deep fakes when they include doctored video images and audio generated with help from AI) to cyberattacks, infringements on data privacy, hiring bias, autonomous vehicles, and drones that do not require a human controller.[80] Similarly, according to a five-country study by KPMG and theUniversity of Queensland Australia in 2021, 66–79% of citizens in each country believe that the impact of AI on society is uncertain and unpredictable; 96% of those surveyed expect AI governance challenges to be managed carefully.[81]
Not only companies, but many other researchers and citizen advocates recommend government regulation as a means of ensuring transparency, and through it, human accountability. This strategy has proven controversial, as some worry that it will slow the rate of innovation. Others argue that regulation leads to systemic stability more able to support innovation in the long term.[82] TheOECD,UN,EU, and many countries are presently working on strategies for regulating AI, and finding appropriate legal frameworks.[83][84][85][2]
On June 26, 2019, the European Commission High-Level Expert Group on Artificial Intelligence (AI HLEG) published its "Policy and investment recommendations for trustworthy Artificial Intelligence".[86] This is the AI HLEG's second deliverable, after the April 2019 publication of the "Ethics Guidelines for Trustworthy AI". The June AI HLEG recommendations cover four principal subjects: humans and society at large, research and academia, the private sector, and the public sector.[87] The European Commission claims that "HLEG's recommendations reflect an appreciation of both the opportunities for AI technologies to drive economic growth, prosperity and innovation, as well as the potential risks involved" and states that the EU aims to lead on the framing of policies governing AI internationally.[88] To prevent harm, in addition to regulation, AI-deploying organizations need to play a central role in creating and deploying trustworthy AI in line with the principles of trustworthy AI, and take accountability to mitigate the risks.[89]
In June 2024, the EU adopted theArtificial Intelligence Act (AI Act).[90] On August 1st 2024, The AI Actentered into force.[91] The rules gradually apply, with the act becoming fully applicable 24 months after entry into force.[90] The AI Act sets rules on providers and users of AI systems.[90] It follows a risk-based approach, where depending on the risk level, AI systems are prohibited or specific requirements need to be met for placing those AI systems on the market and for using them.[91]
AI has been slowly making its presence more known throughout the world, from chatbots that seemingly have answers for every homework question to generative AI that can create a painting about whatever one desires.[23] AI has become increasingly popular in hiring markets, from the ads that target certain people according to what they are looking for to the inspection of applications of potential hires. Events such asCOVID-19 have sped up the adoption of AI programs in the application process, due to more people having to apply electronically, and with this increase in online applicants the use of AI made the process of narrowing down potential employees easier and more efficient. AI has become more prominent as businesses have to keep up with the times and ever-expanding internet. Processing analytics and making decisions becomes much easier with the help of AI.[41] AsTensor Processing Units (TPUs) andgraphics processing units (GPUs) become more powerful, AI capabilities also increase, forcing companies to use it to keep up with the competition. Managing customers' needs and automating many parts of the workplace leads to companies having to spend less money on employees.
AI has also seen increased usage in criminal justice and healthcare. For medicinal means, AI is being used more often to analyze patient data to make predictions about future patients' conditions and possible treatments. These programs are calledclinical decision support systems (DSS). AI's future in healthcare may develop into something further than just recommended treatments, such as referring certain patients over others, leading to the possibility of inequalities.[92]
In 2020, professor Shimon Edelman noted that only a small portion of work in the rapidly growing field of AI ethics addressed the possibility of AIs experiencing suffering. This was despite credible theories having outlined possible ways by which AI systems may become conscious, such as theglobal workspace theory or theintegrated information theory. Edelman notes one exception had beenThomas Metzinger, who in 2018 called for a global moratorium on further work that risked creating conscious AIs. The moratorium was to run to 2050 and could be either extended or repealed early, depending on progress in better understanding the risks and how to mitigate them. Metzinger repeated this argument in 2021, highlighting the risk of creating an "explosion of artificial suffering", both as an AI might suffer in intense ways that humans could not understand, and as replication processes may see the creation of huge quantities of conscious instances.[93][94] Podcast host Dwarkesh Patel said he cared about making sure no "digital equivalent offactory farming" happens.[95] In theethics of uncertain sentience, theprecautionary principle is often invoked.[96]
Several labs have openly stated they are trying to create conscious AIs. There have been reports from those with close access to AIs not openly intended to be self aware, that consciousness may already have unintentionally emerged.[97] These includeOpenAI founderIlya Sutskever in February 2022, when he wrote that today's large neural nets may be "slightly conscious". In November 2022,David Chalmers argued that it was unlikely current large language models likeGPT-3 had experienced consciousness, but also that he considered there to be a serious possibility that large language models may become conscious in the future.[94][93][98]Anthropic hired its first AI welfare researcher in 2024,[99] and in 2025 started a "model welfare" research program that explores topics such as how to assess whether a model deserves moral consideration, potential "signs of distress", and "low-cost" interventions.[100]
According to Carl Shulman andNick Bostrom, it may be possible to create machines that would be "superhumanly efficient at deriving well-being from resources", called "super-beneficiaries". One reason for this is that digital hardware could enable much faster information processing than biological brains, leading to a faster rate ofsubjective experience. These machines could also be engineered to feel intense and positive subjective experience, unaffected by thehedonic treadmill. Shulman and Bostrom caution that failing to appropriately consider the moral claims of digital minds could lead to a moral catastrophe, while uncritically prioritizing them over human interests could be detrimental to humanity.[101][102]
Joseph Weizenbaum[103] argued in 1976 that AI technology should not be used to replace people in positions that require respect and care, such as:
Weizenbaum explains that we require authentic feelings ofempathy from people in these positions. If machines replace them, we will find ourselves alienated, devalued and frustrated, for the artificially intelligent system would not be able to simulate empathy. Artificial intelligence, if used in this way, represents a threat to human dignity. Weizenbaum argues that the fact that we are entertaining the possibility of machines in these positions suggests that we have experienced an "atrophy of the human spirit that comes from thinking of ourselves as computers."[104]
Pamela McCorduck counters that, speaking for women and minorities "I'd rather take my chances with an impartial computer", pointing out that there are conditions where we would prefer to have automated judges and police that have no personal agenda at all.[104] However,Kaplan and Haenlein stress that AI systems are only as smart as the data used to train them since they are, in their essence, nothing more than fancy curve-fitting machines; using AI to support a court ruling can be highly problematic if past rulings show bias toward certain groups since those biases get formalized and ingrained, which makes them even more difficult to spot and fight against.[105]
Weizenbaum was also bothered that AI researchers (and some philosophers) were willing to view the human mind as nothing more than a computer program (a position now known ascomputationalism). To Weizenbaum, these points suggest that AI research devalues human life.[103]
AI founderJohn McCarthy objects to the moralizing tone of Weizenbaum's critique. "When moralizing is both vehement and vague, it invites authoritarian abuse", he writes.Bill Hibbard[106] writes that "Human dignity requires that we strive to remove our ignorance of the nature of existence, and AI is necessary for that striving."
As the widespread use ofautonomous cars becomes increasingly imminent, new challenges raised by fully autonomous vehicles must be addressed.[107][108] There have been debates about the legal liability of the responsible party if these cars get into accidents.[109][110] In one report where a driverless car hit a pedestrian, the driver was inside the car but the controls were fully in the hand of computers. This led to a dilemma over who was at fault for the accident.[111]
In another incident on March 18, 2018,Elaine Herzberg was struck and killed by a self-drivingUber in Arizona. In this case, the automated car was capable of detecting cars and certain obstacles in order to autonomously navigate the roadway, but it could not anticipate a pedestrian in the middle of the road. This raised the question of whether the driver, pedestrian, the car company, or the government should be held responsible for her death.[112]
Currently, self-driving cars are considered semi-autonomous, requiring the driver to pay attention and be prepared to take control if necessary.[113][failed verification] Thus, it falls on governments to regulate drivers who over-rely on autonomous features and to inform them that these are just technologies that, while convenient, are not a complete substitute. Before autonomous cars become widely used, these issues need to be tackled through new policies.[114][115][116]
Experts contend that autonomous vehicles ought to be able to distinguish between rightful and harmful decisions since they have the potential of inflicting harm.[117] The two main approaches proposed to enable smart machines to render moral decisions are the bottom-up approach, which suggests that machines should learn ethical decisions by observing human behavior without the need for formal rules or moral philosophies, and the top-down approach, which involves programming specific ethical principles into the machine's guidance system. However, there are significant challenges facing both strategies: the top-down technique is criticized for its difficulty in preserving certain moral convictions, while the bottom-up strategy is questioned for potentially unethical learning from human activities.
Some experts and academics have questioned the use of robots for military combat, especially when such robots are given some degree of autonomous functions.[118] The US Navy has funded a report which indicates that as military robots become more complex, there should be greater attention to implications of their ability to make autonomous decisions.[119][120] The President of theAssociation for the Advancement of Artificial Intelligence has commissioned a study to look at this issue.[121] They point to programs like the Language Acquisition Device which can emulate human interaction.
On October 31, 2019, the United States Department of Defense's Defense Innovation Board published the draft of a report recommending principles for the ethical use of artificial intelligence by the Department of Defense that would ensure a human operator would always be able to look into the 'black box' and understand the kill-chain process. However, a major concern is how the report will be implemented.[122] The US Navy has funded a report which indicates that asmilitary robots become more complex, there should be greater attention to implications of their ability to make autonomous decisions.[119][120] Some researchers state thatautonomous robots might be more humane, as they could make decisions more effectively.[123] In 2024, theDefense Advanced Research Projects Agency funded a program,Autonomy Standards and Ideals with Military Operational Values (ASIMOV), to develop metrics for evaluating the ethical implications of autonomous weapon systems by testing communities.[124][125]
Research has studied how to make autonomous systems with the ability to learn using assigned moral responsibilities. "The results may be used when designing future military robots, to control unwanted tendencies to assign responsibility to the robots."[126] From aconsequentialist view, there is a chance that robots will develop the ability to make their own logical decisions on whom to kill and that is why there should be a setmoral framework that the AI cannot override.[127]
There has been a recent outcry with regard to the engineering of artificial intelligence weapons that have included ideas of arobot takeover of mankind. AI weapons do present a type of danger different from that of human-controlled weapons. Many governments have begun to fund programs to develop AI weaponry. The United States Navy recently announced plans to developautonomous drone weapons, paralleling similar announcements by Russia and South Korea[128] respectively. Due to the potential of AI weapons becoming more dangerous than human-operated weapons,Stephen Hawking andMax Tegmark signed a "Future of Life" petition[129] to ban AI weapons. The message posted by Hawking and Tegmark states that AI weapons pose an immediate danger and that action is required to avoid catastrophic disasters in the near future.[130]
"If any major military power pushes ahead with the AI weapon development, a globalarms race is virtually inevitable, and the endpoint of this technological trajectory is obvious: autonomous weapons will become theKalashnikovs of tomorrow", says the petition, which includesSkype co-founderJaan Tallinn and MIT professor of linguisticsNoam Chomsky as additional supporters against AI weaponry.[131]
Physicist and Astronomer RoyalSir Martin Rees has warned of catastrophic instances like "dumb robots going rogue or a network that develops a mind of its own."Huw Price, a colleague of Rees at Cambridge, has voiced a similar warning that humans might not survive when intelligence "escapes the constraints of biology". These two professors created theCentre for the Study of Existential Risk at Cambridge University in the hope of avoiding this threat to human existence.[130]
Regarding the potential for smarter-than-human systems to be employed militarily, theOpen Philanthropy Project writes that these scenarios "seem potentially as important as the risks related to loss of control", but research investigating AI's long-run social impact have spent relatively little time on this concern: "this class of scenarios has not been a major focus for the organizations that have been most active in this space, such as theMachine Intelligence Research Institute (MIRI) and theFuture of Humanity Institute (FHI), and there seems to have been less analysis and debate regarding them".[132]
Academic Gao Qiqi writes that military use of AI risks escalating military competition between countries and that the impact of AI in military matters will not be limited to one country but will have spillover effects.[133]: 91 Gao cites the example of U.S. military use of AI, which he contends has been used as a scapegoat to evade accountability for decision-making.[133]: 91
Asummit was held in 2023 in the Hague on the issue of using AI responsibly in the military domain.[134]
Vernor Vinge, among numerous others, has suggested that a moment may come when some or all computers will be smarter than humans. The onset of this event is commonly referred to as "the Singularity"[135] and is the central point of discussion in the philosophy ofSingularitarianism. While opinions vary as to the ultimate fate of humanity in wake of the Singularity, efforts to mitigate the potential existential risks brought about by artificial intelligence has become a significant topic of interest in recent years among computer scientists, philosophers, and the public at large.
Many researchers have argued that, through anintelligence explosion, a self-improving AI could become so powerful that humans would not be able to stop it from achieving its goals.[136] In his paper "Ethical Issues in Advanced Artificial Intelligence" and subsequent bookSuperintelligence: Paths, Dangers, Strategies, philosopherNick Bostrom argues that artificial intelligence has the capability to bring about human extinction. He claims that anartificial superintelligence would be capable of independent initiative and of making its own plans, and may therefore be more appropriately thought of as an autonomous agent. Since artificial intellects need not share our human motivational tendencies, it would be up to the designers of the superintelligence to specify its original motivations. Because a superintelligent AI would be able to bring about almost any possible outcome and to thwart any attempt to prevent the implementation of its goals, many uncontrolledunintended consequences could arise. It could kill off all other agents, persuade them to change their behavior, or block their attempts at interference.[137][138]
However, Bostrom contended that superintelligence also has the potential to solve many difficult problems such as disease, poverty, and environmental destruction, and could helphumans enhance themselves.[139]
Unless moral philosophy provides us with a flawless ethical theory, an AI's utility function could allow for many potentially harmful scenarios that conform with a given ethical framework but not "common sense". According toEliezer Yudkowsky, there is little reason to suppose that an artificially designed mind would have such an adaptation.[140] AI researchers such asStuart J. Russell,[141]Bill Hibbard,[106]Roman Yampolskiy,[142]Shannon Vallor,[143]Steven Umbrello[144] andLuciano Floridi[145] have proposed design strategies for developing beneficial machines.
To address ethical challenges in artificial intelligence, developers have introduced various systems designed to ensure responsible AI behavior. Examples includeNvidia'sLlama Guard, which focuses on improving thesafety andalignment of large AI models,[146] andPreamble's customizable guardrail platform.[147] These systems aim to address issues such as algorithmic bias, misuse, and vulnerabilities, includingprompt injection attacks, by embedding ethical guidelines into the functionality of AI models.
Prompt injection, a technique by which malicious inputs can cause AI systems to produce unintended or harmful outputs, has been a focus of these developments. Some approaches use customizable policies and rules to analyze inputs and outputs, ensuring that potentially problematic interactions are filtered or mitigated.[147] Other tools focus on applying structured constraints to inputs, restricting outputs to predefined parameters,[148] or leveraging real-time monitoring mechanisms to identify and address vulnerabilities.[149] These efforts reflect a broader trend in ensuring that artificial intelligence systems are designed with safety and ethical considerations at the forefront, particularly as their use becomes increasingly widespread in critical applications.[150][151]
There are many organizations concerned with AI ethics and policy, public and governmental as well as corporate and societal.
Amazon,Google,Facebook,IBM, andMicrosoft have established anon-profit, The Partnership on AI to Benefit People and Society, to formulate best practices on artificial intelligence technologies, advance the public's understanding, and to serve as a platform about artificial intelligence. Apple joined in January 2017. The corporate members will make financial and research contributions to the group, while engaging with the scientific community to bring academics onto the board.[152]
TheIEEE put together a Global Initiative on Ethics of Autonomous and Intelligent Systems which has been creating and revising guidelines with the help of public input, and accepts as members many professionals from within and without its organization. The IEEE'sEthics of Autonomous Systems initiative aims to address ethical dilemmas related to decision-making and the impact on society while developing guidelines for the development and use of autonomous systems. In particular, in domains like artificial intelligence and robotics, the Foundation for Responsible Robotics is dedicated to promoting moral behavior as well as responsible robot design and use, ensuring that robots maintain moral principles and are congruent with human values.
Traditionally,government has been used by societies to ensure ethics are observed through legislation and policing. There are now many efforts by national governments, as well as transnational government andnon-government organizations to ensure AI is ethically applied.
AI ethics work is structured by personal values and professional commitments, and involves constructing contextual meaning through data and algorithms. Therefore, AI ethics work needs to be incentivized.[153]
Historically speaking, the investigation of moral and ethical implications of "thinking machines" goes back at least to theEnlightenment:Leibniz already posed the question of whether we should attribute intelligence to a mechanism that behaves as if it were a sentient being,[181] and so doesDescartes, who describes what could be considered an early version of theTuring test.[182]
Theromantic period has several times envisioned artificial creatures that escape the control of their creator with dire consequences, most famously inMary Shelley'sFrankenstein. The widespread preoccupation with industrialization and mechanization in the 19th and early 20th century, however, brought ethical implications of unhinged technical developments to the forefront of fiction:R.U.R – Rossum's Universal Robots,Karel Čapek's play of sentient robots endowed with emotions used as slave labor is not only credited with the invention of the term 'robot' (derived from the Czech word for forced labor,robota)[183] but was also an international success after it premiered in 1921.George Bernard Shaw's playBack to Methuselah, published in 1921, questions at one point the validity of thinking machines that act like humans;Fritz Lang's 1927 filmMetropolis shows anandroid leading the uprising of the exploited masses against the oppressive regime of atechnocratic society.In the 1950s,Isaac Asimov considered the issue of how to control machines inI, Robot. At the insistence of his editorJohn W. Campbell Jr., he proposed theThree Laws of Robotics to govern artificially intelligent systems. Much of his work was then spent testing the boundaries of his three laws to see where they would break down, or where they would create paradoxical or unanticipated behavior.[184] His work suggests that no set of fixed laws can sufficiently anticipate all possible circumstances.[185] More recently, academics and many governments have challenged the idea that AI can itself be held accountable.[186] A panel convened by theUnited Kingdom in 2010 revised Asimov's laws to clarify that AI is the responsibility either of its manufacturers, or of its owner/operator.[187]
Eliezer Yudkowsky, from theMachine Intelligence Research Institute, suggested in 2004 a need to study how to build a "Friendly AI", meaning that there should also be efforts to make AI intrinsically friendly and humane.[188]
In 2009, academics and technical experts attended a conference organized by theAssociation for the Advancement of Artificial Intelligence to discuss the potential impact of robots and computers, and the impact of the hypothetical possibility that they could become self-sufficient and make their own decisions. They discussed the possibility and the extent to which computers and robots might be able to acquire any level of autonomy, and to what degree they could use such abilities to possibly pose any threat or hazard.[189] They noted that some machines have acquired various forms of semi-autonomy, including being able to find power sources on their own and being able to independently choose targets to attack with weapons. They also noted that some computer viruses can evade elimination and have achieved "cockroach intelligence". They noted that self-awareness as depicted in science-fiction is probably unlikely, but that there were other potential hazards and pitfalls.[135]
Also in 2009, during an experiment at the Laboratory of Intelligent Systems in the Ecole Polytechnique Fédérale ofLausanne, Switzerland, robots that were programmed to cooperate with each other (in searching out a beneficial resource and avoiding a poisonous one) eventually learned to lie to each other in an attempt to hoard the beneficial resource.[190]
The role of fiction with regards to AI ethics has been a complex one.[191] One can distinguish three levels at which fiction has impacted the development of artificial intelligence and robotics: Historically, fiction has prefigured common tropes that have not only influenced goals and visions for AI, but also outlined ethical questions and common fears associated with it. During the second half of the twentieth and the first decades of the twenty-first century, popular culture, in particular movies, TV series and video games have frequently echoed preoccupations and dystopian projections around ethical questions concerning AI and robotics. Recently, these themes have also been increasingly treated in literature beyond the realm of science fiction. And, as Carme Torras, research professor at theInstitut de Robòtica i Informàtica Industrial (Institute of robotics and industrial computing) at the Technical University of Catalonia notes,[192] in higher education, science fiction is also increasingly used for teaching technology-related ethical issues in technological degrees.
While ethical questions linked to AI have been featured in science fiction literature andfeature films for decades, the emergence of the TV series as a genre allowing for longer and more complex story lines and character development has led to some significant contributions that deal with ethical implications of technology. The Swedish seriesReal Humans (2012–2013) tackled the complex ethical and social consequences linked to the integration of artificial sentient beings in society. The British dystopian science fiction anthology seriesBlack Mirror (2013–Present) is particularly notable for experimenting with dystopian fictional developments linked to a wide variety of recent technology developments. Both the French seriesOsmosis (2020) and British seriesThe One deal with the question of what can happen if technology tries to find the ideal partner for a person. Several episodes of the Netflix seriesLove, Death+Robots have imagined scenes of robots and humans living together. The most representative one of them is S02 E01, which shows how bad the consequences can be when robots get out of control if humans rely too much on them in their lives.[193]
The movieThe Thirteenth Floor suggests a future wheresimulated worlds with sentient inhabitants are created by computergame consoles for the purpose of entertainment. The movieThe Matrix suggests a future where the dominant species on planet Earth are sentient machines and humanity is treated with utmostspeciesism. The short story "The Planck Dive" suggests a future where humanity has turned itself into software that can be duplicated and optimized and the relevant distinction between types of software is sentient and non-sentient. The same idea can be found in theEmergency Medical Hologram ofStarship Voyager, which is an apparently sentient copy of a reduced subset of the consciousness of its creator,Dr. Zimmerman, who, for the best motives, has created the system to give medical assistance in case of emergencies. The moviesBicentennial Man andA.I. deal with the possibility of sentient robots that could love.I, Robot explored some aspects of Asimov's three laws. All these scenarios try to foresee possibly unethical consequences of the creation of sentient computers.[194]
Over time, debates have tended to focus less and less onpossibility and more ondesirability,[195] as emphasized in the"Cosmist" and "Terran" debates initiated byHugo de Garis andKevin Warwick.
{{citation}}: CS1 maint: work parameter with ISBN (link){{citation}}: CS1 maint: work parameter with ISBN (link){{citation}}: CS1 maint: work parameter with ISBN (link){{citation}}: CS1 maint: work parameter with ISBN (link)| Concepts |
|
|---|---|
| Organizations |
|
| People |
|
| Books | |
| Other | |
| Concepts |
| ||
|---|---|---|---|
| Theories |
| ||
| Philosophy of... | |||
| Related topics |
| ||
| Philosophers of science |
| ||
