OpenAI and DeepMind both have a focus on safety and both have safety teams.

Judged as of 2017-07-21. True: dated sources show that by July 2017 both OpenAI and DeepMind had a stated focus on AI safety and that each had a dedicated safety team or group, with named staff working on AI safety research. Mostly true: both had a clear safety focus and dedicated safety researchers, but one organisation's group was not formally called a team or was very small. Mixed: only one of the two organisations had a safety team at that time, though both stated a safety focus. Mostly false: both stated concern for safety, but neither had an identifiable safety team in July 2017. False: evidence shows that neither had a safety focus or safety staff at that time.

Set before the research began.

1. E1 Learning through human feedback DeepMind · 12 June 2017

   “That is why DeepMind co-founded initiatives like the Partnership on AI to Benefit People and Societyand why we have a team dedicated to technical AI Safety.”

2. E2 Learning from human preferences OpenAI · 13 June 2017 Own record

   “This post is representative of the work done by OpenAI’s safety team; if you’re interested in working on problems like this, please join us!”

3. E3 Learning through human feedback DeepMind · 12 June 2017

   “This research was done as part of an ongoing collaboration between Jan Leike, Miljan Martic, and Shane Legg at DeepMind and Paul Christiano, Dario Amodei, and Tom Brown at OpenAI.”

4. E4 #3 - Dario Amodei on OpenAI and how AI will change the world for good and ill Apple Podcasts (podcast episode listing) · 21 July 2017

   “His paper *Concrete Problems in AI Safety*, which outlines five specific ways machine learning algorithms can act in dangerous ways their designers don’t intend - something OpenAI has to work to avoid.”

DeepMind's blog dated 2017-06-12 says DeepMind has "a team dedicated to technical AI Safety" (E1) and names Jan Leike, Miljan Martic and Shane Legg as its researchers on the joint safety work (E3). OpenAI's official post dated 2017-06-13 calls the work "representative of the work done by OpenAI's safety team" (E2), and E3 names Paul Christiano, Dario Amodei and Tom Brown as the OpenAI researchers. E4 also describes OpenAI working to avoid AI safety problems. By July 2017, both organisations had a stated safety focus, a dedicated safety team and named safety staff, which meets the criteria for true.