I am not a robot. You are not a chatbot. But we may soon all have to work harder to prove it. We appear to be at a critical phase in the regulation of the development of artificial intelligence (AI). In January IBM warned of AI replacing about 30% of their planned recruitment in back-office functions. In March 600 leading figures in the tech sector signed an open letter calling for a six-month moratorium on the development of the most powerful AI systems. The head of compliance for Google Geoffrey Hinton has just resigned his post to give himself the freedom to speak out against "bad actors" - developers who build systems capable of confusing fact with fiction.
The Biden administration has reacted; on April 25th four key regulatory bodies including the Equal Employment Opportunity Commission (EEOC) issued a joint statement under the chair of the EEOC Lina Khan, warning that AI did not provide a cover for breaking laws on safeguarding competition or anti-discrimination. The EEOC has a general concern that the AI technology can and will be used to filter out older and disabled workers from the workforce.
Among the examples given of popular work-related AI tools are resume scanners, employee monitoring software that ranks employees based on keystrokes, game like on-line tests to assess job skills and video software that measures speech patterns or facial expressions. Technology used in these contexts is often not programmed with legally permitted reasonable adjustments such as a person with autism who requires a quieter workstation or a person with a pregnancy related health condition who requires more comfort breaks.
So, what has the UK Government had to say about US developments that are concurrently happening in our own country too? At the end of March, the Government issued a white paper on AI. It contained four mentions of the word discrimination. These mentions tended to contextualise discrimination as a risk to public trust and confidence in AI, not the possibility of AI, undermining confidence in equalities. On April 24th the Prime Minister sought to copy the vaccines task force and announced £100 million investment in AI. The Prime Minister spoke glowingly of the need to advance AI foundation models in safe and reliable ways but offered no detail on safeguards over anti-discrimination and protecting equalities. On the basis of government statements, the UK Government appears to regard rights as requiring navigation whereas the US Government requires them to be upheld.
Now is the time to exercise moral judgement on the technology before the technology itself becomes our master. We must have critics and people with equalities and legal expertise on the task force. If you have reservations on AI write to the Prime Minister and cc in the Science Minister, Michellle Donnellan. Please respond to the white paper on AI and do it now. Please copy the Disability Resilience Network in too. We are not averse to scientific progress, but we want responsive ministers investing in responsible science.
Comments