What a way to end the week with a great @OECD session on #AIEthics in the workplace with two of my favourite people @CjColclough and @mer__edith π€© #oecdAI Thread here : (1/x) 
The core issue of #AIEthics at work is the fact algorithms are βbosses tools β used by people who already have power - leadership - over disempowered - employees when we have so little safe guards & protection says @mer__edith ππ»ππ»ππ»And this is a HUGE challenge to overcome (2/x)
Next lovely @CjColclough talks about other 2 main challenges
1. Increased dependency on suppliers of AI that will protect the IP but be less willing to audit and correct their algos for ethical issues
2. Lack of governance of algorithms when in use π€―
#AIethics (3/x)
1. Increased dependency on suppliers of AI that will protect the IP but be less willing to audit and correct their algos for ethical issues
2. Lack of governance of algorithms when in use π€―
#AIethics (3/x)
Disagree with @valeriodeste when he claims #AIethics is not enough bc is so vague, but talk regulation.
Why
1. the term was abused by non experts & as a result has this reputation
2. regulation is routed in ethics
3. Fairness, privacy, accountability are ALL ethical values (4/x)
Why
1. the term was abused by non experts & as a result has this reputation
2. regulation is routed in ethics
3. Fairness, privacy, accountability are ALL ethical values (4/x)
Both @mer__edith and @CjColclough call productivity a faulty goal for AI as is established part of the eternal dream of capitalism - infinite growth. What we need is a different table, where those impacted by AI can seat as equals with leaders to decide AI goals #AIEthics (5/x)
Honestly, this is such a great panel it needs a replay #OECDAI! @valeriodeste talks about the trap of snake oil AI for SMEs. That is why we need proper AI procurement framework. @wef ones is a good start, but more needed @johnchavens @ruchowdh @mona_sloane (6/x) #AIethics
Write this on a T shirt and wear it at work β Gender discrimination wasnβt invented with AI. β @mer__edith ππ»ππ»ππ» only understanding the source we can create a proper solution ( right AI looks like a acceleration of a problem rather than a solution) (7/x)
Throughout Sylvain Duranton @BCG stressed many times the need for a proper governance of AI multilayered, to insure accountability and transparency. @mer__edith reiterated too.
Here is the model I use with my clients. Good governance is solution to many evils #AIEthics ( 8/x)
Here is the model I use with my clients. Good governance is solution to many evils #AIEthics ( 8/x)
β’ β’ β’
Missing some Tweet in this thread? You can try to
force a refresh
γ
