Humanity quietly removed some of the voluntary commitments it made in conjunction with the Biden administration in 2023 from its website to promote safe and “trustworthy” AI.
According to the AI WatchDog Group The Midas Project, commitments that include sharing research into AI risks and AI bias and discrimination across industry and governments were removed from the Transparency Hub for Humanity last week. Other Biden-era commitments relating to AI-generated image-based sexual abuse remain.
It appears that humanity has not notified of the change. The company did not immediately respond to requests for comment.
Humanity, together with companies such as Openai, Google, Microsoft, Meta, and fenction, announced in July 2023 that it had agreed to comply with the specific voluntary AI safety commitments proposed by the Biden administration. The commitments included internal and external security testing of AI systems prior to release, investments in cybersecurity to protect sensitive AI data, and development of watermarks for AI-generated content.
To be clear, humanity has already adopted many practices outlined in the commitment, and consent was not legally binding. However, the Biden administration’s intention was to signal the prioritization of AI policy ahead of a more thorough AI executive order that came into effect several months later.
The Trump administration shows that its approach to AI governance is completely different.
In January, President Trump abolished the aforementioned AI executive order. This identified defects in the model that contained bias and directed the author’s guidance to help them do the right thing. Critics who allied with Trump argued that the reporting requirements for the order were tedious and that businesses were forced to disclose their trade secrets.
Shortly after rescinding the AI executive order, Trump signed an order directing federal agencies to promote the “ideologically unbiased” development of AI that promotes “human prosperity, economic competition and national security.” Importantly, Trump’s order did not mention the key tenet of Biden’s initiative, the fight against AI discrimination.
As the MIDAS Project pointed out in a series of posts on X, nothing of Biden-era commitment was suggested. In November, several AI companies confirmed their commitments had not changed since the election.
Humanity is not the only company to coordinate public policy in the months since Trump took office. Openai recently adopted “Intellectual freedom… no matter how challenging or controversial the topic” and announced that it would strive to prevent its AI from censoring certain perspectives.
Openai also scrubed pages on websites used to express the startup’s commitment to diversity, equity, inclusion, or DEI. These programs have been attacked by the Trump administration, with many businesses eliminating or effectively reviving the DEI initiative.
Many Trump’s Silicon Valley advisors on AI, including Mark Andreesen, David Sachs and Elon Musk, claim that companies, including Google and Openai, are engaged in AI censorship by limiting the answers to AI Chatbots. Laboratories, including Openai, have denied that policy changes are responding to political pressures.
Both Openai and humanity are actively pursuing government contracts.
Hours after the story was released, humanity sent the following statement to TechCrunch:
“We continue to be committed to the voluntary AI commitments established under the Biden administration, and this progress and certain actions continue to reflect. [our] Transparency center within content. To prevent further confusion, add a section that directly cites where progress is consistent. ”
Updated 11:25am Pacific: Added statement of humanity.
Source link
