OpenAI, the artificial intelligence research lab known for its powerful language models, has backtracked on a controversial policy that forced departing employees to choose between speaking out about the company and receiving their earned financial compensation. This move comes after a recent report exposed the harsh terms of OpenAI’s nondisparagement agreements, sparking criticism and a potential exodus of disgruntled employees.

According to Bloomberg, OpenAI previously presented exiting employees with a stark choice: sign a legal document promising not to criticize the company or forfeit their vested stock options, which in some cases could amount to millions of dollars. This essentially forced employees to remain silent about any negative experiences or concerns they might have had at OpenAI, raising serious ethical questions about free speech and employee rights.

The situation took a turn for the worse when Vox revealed the existence of this policy. OpenAI CEO Sam Altman, seemingly unaware of the provision’s inclusion, expressed his embarrassment and vowed to have the company’s exit paperwork revised. The company has since confirmed that it will not enforce any existing nondisparagement agreements and will completely remove such language from future documents.

While this is a positive step towards a more transparent and employee-friendly environment, the damage may already be done. Reports suggest that at least one former employee lost a significant amount of vested equity by refusing to sign the agreement. OpenAI has remained silent on whether this individual will be compensated for their lost earnings.

The company has also reached out to current employees to alleviate concerns about potential repercussions for speaking their minds. Chief Strategy Officer Jason Kwon issued a statement apologizing for the “distress this has caused” and promised to work diligently towards creating a better work environment.

This recent controversy is just one of several issues OpenAI has faced lately. The company recently disbanded its team dedicated to ensuring the safe development of artificial intelligence, raising concerns about OpenAI’s commitment to responsible AI research. Additionally, the departure of key personnel like Chief Scientist Ilya Sutskever and accusations of misappropriating Scarlett Johansson’s voice for a virtual assistant cast a shadow over OpenAI’s reputation.

OpenAI’s decision to scrap its nondisparagement agreements is a step in the right direction. However, the company has a lot of ground to cover in regaining the trust of its employees and the wider AI research community.