The flip side of AI
Biased decisions taken by AI have been a recurring threat to industries deploying it and to the ethical norms of the society at large. While unbalanced decisions affect business outcomes, biases related to gender and race negatively impacts the social fabric. Expert AI luminaries are anxious about the biasness apparent in these machine–learning algorithms that take millions of decisions every minute.
AI-enabled machines reflect a persona basis the kind of data that they are fed. As a result, they acquire biases from the datasets that they are fed and cannot be immune to human biases. No wonder, biased data will make them take biased decisions. Combating biases, both human and data-based, is one of the biggest ethical challenges that AI is faced with. What’s more alarming is that the problem is expected to become more severe in the near future. This is because many are yet to realize the magnitude of the impact that feeding biased data will have on businesses. Therefore, steps are yet to be taken to identify solutions to the compelling problem.
The need for an open discussion
NASSCOM Product Conclave took the initiative to launch a platform for the industry to openly discuss different opinions on the subject. We planned a Twitter Campaign with the hashtag #DeconstructingAI, on 31st January 2018 at 4 P.M. on our Twitter handle, @NASSCOM_Product. The purpose was to discuss our polarising ideas and see the subject from different points of view and attempt to have Tech experts bust some myths and perceptions.
The debate was centred on 5 key questions.
- Do you believe that intelligent systems will learn human prejudices and how?
- How can we be transparent about the training data we are using, while working on such algorithms?
- How can we prepare ourselves for a bias-free AI?
- Can AI help people in making less biased decisions?
- How do you think Artificial Intelligence will replace human jobs in the future?
The industry responded and it was Overwhelming!
Industry experts including Sridhar Iyengar (European Explorer, Zoho &Product Council Member, NASSCOM), Sangeeta Gupta (Director, NASSCOM), Subinder Khurana (Product Council Member, NASSCOM), Atul Batra (CTO, Manthan & Product Council Member, NASSCOM), Praveen Nallapaneni (Deputy Manager, NASSCOM) and Srikanth Srinivasan (Regional Director TS & AP NASSCOM) joined in the conversations.
The Tweet Chat on #DeconstructingAI was a great success generating over 200 responses within 30 minutes of the chat and driving over 802K impressions on Twitter alone. We’re eagerly looking forward to the next one!
If you could not join the tweetchat and would like to share your responses, comment below.