Building a Winning AI Strategy with Trustworthy AI

A good AI strategy doesn't just verify the value created by a business or support decisions. It can help reveal key metrics that help leaders in determining "are we measuring the metrics driving value creation in our business?"

A winning AI Strategy goes beyond a good AI strategy and De-Risks adoption of AI by identifying risks and laying a foundation of Trustworthy AI. I will give some key scenarios where these Trustworthy pillars of Security, Privacy, Ethics, Transparency, Regulations, Audit and Accountability can be decisive between success and failure.

To adopt AI or not to adopt AI - This is a dilemma for businesses today and one that can have significant impact on their ability to act on Key Performance Indicators, KPIs, Enhanced KPIs and New KPIs that are not currently examined or possible without AI but can add significantly to :

Business innovation and/or growth.

Competitive edge

Talent acquisition and retention

Customer acquisition and satisfaction

Profitability and

Sustainability.

De-Risking AI Strategy

1. Decouple Business risk from Regulation risk.

Companies are in a wait and see mode as they wait for regulations such as EU AI Act to finalize as they're worried it will impact their AI trajectory. That is a mistake as the risk profiles laid out do not correlate and align completely with Business Risk categories and the categories are not always overlapping. ?The EU AI Act sets out a risk-based approach, where?the obligations for a system are proportionate to the level of risk that it poses. The Act outlines four levels of risk:

• High-risk systems - systems that can have a significant impact on the life chances of a user. There are eight types of systems that fall into this category. These systems are subject to stringent obligations and must undergo conformity assessments before being put on the EU market.
• Systems with unacceptable risk - these systems are not permitted to be sold on the EU market.
• Low-risk systems.
• Limited or minimal risk systems

2. Building a “risk-based” Strategy for AI is essential for businesses today. Furthermore, building it with pillars of Trustworthy AI will reduce risk and help create reassurance in these high impact systems.

Scenario 1: Psychological Targeting :

What is it? The practice of influencing behavior through various tactics based on personality, psychological and behavioral traits.

Is it useful for business? What type of business?

As per domo.com in 1 minute

Amazon customers spend $283,000, Facebook receives 44 million views, Instagram users share 65,000 photos and Venmo hosts transactions worth $304,000

What are the risks? It can help predict consumer behavior and preferences, choices consumers are making and also help predict their personality, their motivations potentially. As personalization products ramp up in offerings and functionality this area will be extremely valuable in taking disparate customer data from various media and sources to provide a better customer experience.

If it offers such immense value how do I address the risks?

Companies that are strategic and not reactive would also ask: Even if its not regulated, what are the risks and how can we mitigate them?

Risk based AI Strategy for Psychological targeting would have very important Trustworthy AI components of Ethics and Transparency.

I have covered a scenario here, there are many others which will carry different risks for the business, business ecosystem and society at large.

One point is very important to note: I would like to share what I have learned from twenty years of creating risk based holistic programs for global large companies, some of them in highly complex regulated environments and immense operational complexity -

#1 We do not have the luxury of wait and see or taking a piece meal and reactive approach to managing AI in our business, society or universally.

#2 The reason I have been a crusader for Trustworthy AI for that last 4 years and continue on that path is it can help steer us on to the right course and avoid foreseeable catastrophic and high impact risks.

Part 2 of this article...stay tuned Regulations that mandate Risk Management:

Please post your comments, questions, thoughts below. If you would like to contact me directly, email me at [email protected].