Every software company is clamoring for AI today. SAS is doing so too, but at the same time stresses that reliability is important to adopt the technology.

During SAS Innovate on Tour in Rotterdam, the session by Sterlina Smith, Senior Manager SAS Data Ethics Practice, was well attended. The session revolves around delivering trustworthy AI and it starts with a clear problem: “New innovations always collide with old inequalities.”

Smith says we have two options from there. “We can allow these problems to persist, or we can make changes that limit or mitigate the problem. And this is where we see the AI ​​trust gap.”

“Distrust is at the root of the social, legal, ethical and economic concerns surrounding AI. People find it hard to believe that AI will keep them safe and that it was developed with an understanding of their needs. They doubt that it will benefit them.”

Electricity and electric shock

Every technology brings its own negativity, and AI is no different. Smith brings in French philosopher Paul Virilio. “He who invents the ship also invented the shipwreck. Electricity also brings electric shocks.” She emphasizes that something that grows quickly breeds distrust.

AI fear is a problem today, she says. She cites Edelman’s 2024 Trust Barometer (PDF). The technology is at an important crossroads. “Government and media can help portray AI in a positive light, but that’s also the problem.” She brings up Edelman’s research again.

“This is where we see how organizations can play an important role. Today, people trust companies in their ability to integrate innovation into society. In this way, companies can be a catalyst for the integration and adoption of AI.”

At the same time, Smith stresses that artificial intelligence brings with it great responsibility. “The gap between the richest and the least developed countries must not widen. In the 48 years of our existence, we have learned that we have the greatest impact when we put people at the center.”

transparency

During her presentation, Smith often focuses on synthetic data. “Some data is underrepresented, which makes the datasets not entirely fair. Synthetic data can further solve this problem. Will people believe what they see? How do we combat disinformation? That’s why we’re talking to governments and industry today to advocate for watermarking and labeling. That way, users can clearly see what is authentic and what comes from SAS Data Maker, for example.”

When it comes to AI, you have to involve people.

Sterlina Smith, Senior Manager SAS Data Ethics Practice

It touches on the biggest problem of all when it comes to AI: transparency. At SAS, the techniques are being further improved to find out where the data comes from. “This is necessary for transparency and fairness,” says Smith.

“You have to involve people when it comes to AI. Artificial intelligence can be fantastic, but people have to take action against it.” On the human side, she emphasizes very strongly that there is a conscious approach to innovation. “Trustworthy AI starts with the first line of code.”

Ongoing AI race

The AI ​​race is in full swing and SAS is running with it. Smith is all too aware of this, but she puts it in perspective. “When it comes to ethics, trust and responsible innovation, it’s easy to go down the path of doom. We all look together at what terrible things can happen. But instead of that dark side, I want you to think about the side where the light shines, where things are good? And what happens if we get this right?”

“What if we could target government services to those who need them while preventing fraud and abuse? What if we could reduce the number of police calls and still ensure safety? And what if AI strengthens democracy rather than weakening it? These goals are worth pursuing. I hope you agree with me.”

———————————————————

This is an editorial contribution in collaboration with SAS. For more information about privacy, security and reliability and how SAS handles them, please see here Here.