Deloitte’s Beena Ammanath: Do Workers Trust AI?

hi I'm James Maguire and on today's webcast we're looking at a new report called the state of ethics and Trust in technology which provides a portrait of user attitude about the technology they use including generative AI to discuss that I'm joined by beina amth the leader of Deo technology trust ethics practice beina very good to have you with us today James thank you so much for having me and I'm looking forward to our conversation so uh the annual report state of ethic and Trust in technology fascinating stuff let's talk about some of the individual findings but first what what findings stood out for you overall in the report what really caught your eye well you know there were two things that really caught my eye you know I've been talking about uh trust and ethics and Technology I'm a technologist by training but I've been talking about the need for ethics and trust and technology for a long time and you know I because I truly believe that to reach the full potential of any technology today we are in the AI world to reach the full potential of any technology you should be able to trust it right so trust is a crucial factor and what I've seen is that same level of pragmatic optimism come out in this survey findings there's obviously a lot of concern about the risk the you know trust and ethics factors but also there's a lot of optimism on how this technology AI can actually be used for a lot of social good so you know I love that the survey findings actually resonated on that factor the other one that resonated with me was how many comp organizations have introduced ethical Tech trainings in their organizations but the biggest wow for me was it actually changed the behavior after the people took the training right which is which I think is phenomenal that you know the training is having that level of impact is is the ethical training is it about avoiding bias is about you know security compliance with the technology what is the main focus of the ethics training so you know I think most of the ethical tech training talks about the newer risks that are introduced uh with with the technology like AI right the traditional Technologies we've been through the internet era the cloud era there there there are those those risk as well but AI introduces nuances around it plus the new risks that come with it like bias the need for fa fairness the need for transparency um you know we obviously you know we've all heard about hallucination which actually in my I think in my world I'm thinking of it is it's more of a reliability issue we call it a cute little world word called but reliability issues have always been there in the software world so it's looking at Safety and Security compliance and need for governance and controls so all of that is covered in addition to each in organization talking about the specific internal AI tools that's available to the employees educating the employees on what tools are available what's the you know who can they reach out to if they have concerns about the technology or if they notice certain Behavior so it's it's both a a conceptual plus also grounded in what the organization what the resources are available in the organization interesting okay well all right one more thing before we look at the findings I mean do you have a global view of how organizations are navigating the challenges of generative AI it seems like there's so much potential there but there is a sense of a liness like we don't know we we we're inspired but we're also a little bit scared of it um correct me if I'm wrong on that score but like how are companies handling it in your sense yeah this was a global report uh you know with over 1,800 uh respondents and you know I think um there is uh obviously the uh we seeing um concerns around regulations and what might come uh you know the every region is in a different place right now and the survey reflects that the uh so we seeing the need for regulations the awareness that upcoming regulations are there but also there is an equal optimism on how can this technology be harnessed into driving real social impact driving real business value so there's a good level of balanced appro uh Viewpoint that the survey reveals H okay all right let's look at a few key findings uh let's talk through these Safety First all in this year's survey 78% of respondents selected safe and secure as one of the top three ethical technology principles a 37% increase from respondents in the previous year's survey what how do you take that point yeah I think you know uh and I James I speak to a lot of boards and CEOs and uh what I've seen is you we are seeing this evolving into where you're looking at trust and ethics pretty much like from a c you know when cyber was new and when it became a board topic and when there were robust rules set up right so safety and secure is kind of building on top of what cyber has done and elevating the role of cyber the need for cyber security in this new AI era and cyber is Something board members understand most of us understand because it's been the concerns around that have been there longer so safety and secure is kind of how you look at the nuances that AI brings in to continue to apply the same cyber security principles and what are the new challenges that that cyber will have to face so we we are seeing that increase and it's it's more of U I think it's uh accessible to a wider audience the importance of cyber and we see you know Ai and the impact on Cyber rising to the top well and this maybe a bit of a side point but it seems like on on the question of cyber AI is such a two-sided coin and that you know the good actors have it the companies have it but also the the Bad actors the malicious folks have it as well and so it's a bit of a a stand UPF I'm not sure who's going to win that battle probably wasn't covered in your survey but I think that does it does influence the issue of trust yes yes absolutely and look today you know any bad actor with an internet connection can do it because we've entered the era of AI where the technology is changing and evolving you look at technology 25 years ago you know you deployed a piece of software it you know it was it stay the outcome stayed you know you put in a specific input the output came out the same every single time right now this technology AI is still learning and evolving after it is deployed so you know it is um it is one of those areas where Bad actors can actually morph change the technology and we've seen it where chat Bots go racist or misogynistic where you know B actors all that they need is an internet connection to start making the technology fail all right here's here's another point from the survey uh organizations face challenges building trust among employees trust in one's organization and its emerging Technologies may be declining pronounced in younger Generations so there how do you see it seems like that the younger folks AR aren't as trusting uh it's it's an interesting finding right and I think part of it is you know in 2023 87% of Millennials and 81% of gen zer respondents reported by into to the organization's ethical messaging what happened this year is those figures decreased to 77% and 65% respectively right and overall it fell from 89 to 81 but in the younger generation there's almost that decline in trust uh maybe because they are not seeing the messaging that was communicated aligning with some of the questions that's coming up to us coming up to the organization so to really help build trust organizations have to be transparent have to make sure there are these trainings available the employees are fully familiar with all the controls that are in place for employees to access I think uh you're seeing that generational shift in trust but there are way the good news is there are ways to mediate it such as what what might be one way to mediate it uh one is transparency if you have controls in place if you have checks and balances if you're using say if you're using AI in your HR System right be transparent about it that we use AI to do the first level of passing through new resumés that we get or if you're using AI to write up that you know performance review make sure you're transparent that the first draft will be generated by AI make sure the employees are familiar with the tool make sure the employees have a way to get engaged and provide feedback on the tools I mean HR was just one of the example but really in any function or and and especially the function that the employee might be working in to be super transparent about what when and where AI is being used what kind of checks are being done providing mechanisms maybe it's that 1 1800 number or an email where employees can reach out to if they have concerns about a specific AI tool that is that they're using on a regular basis in the uh in the uh organization so transparency training uh making employees part of the journey as opposed to just put pushing out the software I think those are a few ways um to start building out that trust AI might be a technology but you and I both know it's not a technology Pro just a technology project you know success of it depends on a cultural change uh there is a human impact and there is you know there is real concerns about you know the jobs getting impacted so being proactive and addressing it up front um uh is is is one of the you know crucial ways to unback trust it it's true you said about AI it's obviously it's technology and it's it's a huge project it's a huge initiative I think it's almost something emotional in people's attitude towards it there is a threat of it is it going to take my job or what what is it can replace me some so it's like it's it's it's not like you know cloud or data analytics it's a completely new technology in that sense yes and you know anything that comes um close to human intelligence and the there is going to be that concern on you know oh can technology can AI take over the the job that I'm doing and the reality is you know uh the jobs are going to change right I I like to say that it's not uh you know it's not AI going to take over your job it's the person it's a human that knows AI right that will take over your job the human being the employee that knows how to use AI tools most effectively in their job it's like it's like you know you have to learn and evolve and the more organizations can provide those AI fluency trainings or AI literacy training make sure employees are well trained the more that builds trust as well all right another another survey another uh from the finding from the survey is uh it says reputation is top of mind when asked to rank rank the potential negative outcomes to organizations if ethical standards are not followed uh for emerging Technologies respondents rated reputational damage 82% Financial damage 66% and Regulatory penalties 60 60% as their top three concerns yeah yeah and two I think part of it is because there have been so many you know news headlines about you know um technology going wrong uh there have been lawsuits and all of it comes down to reputation of the organization I was pleasantly surprised that you know most most people are aware that is the biggest that is the biggest damage that could happen uh you know the reputation that you build and it's not only from an external perspective even on to your own employees you know if there is a headline about your AI going rogue then there is there is a real risk for for both internal and external and it can set back the entire AI journey of an organization back sure I mean we remember there there's a large AI provider whose name I won't mention that they they debuted a a new tool and we had high high hopes and that tool was sort of you know it it did not behave properly and it was extremely embarrassing for that provider I thought oh oh goodness they should have taken a little bit more time with that tool before they debuted it and we'll see more of those uh you know I can't speak to specific companies organizations but you know I I think um I think this survey actually shows the need to do that that level of testing and you know I don't know if you noticed this James but you know one of the key findings was also you know how organizations that are uh engaging in focus groups getting diverse focus groups are you know make are able to make their AI more robust make it more you know it it the more diversity and F uh you know you can bring in those external perspectives it really helps make the AI more trustworthy you mean even as ai ai is being developed those diverse voices are involved exactly yes yes I mean that's a topic close to my heart but you know I think diver diversity of needs to be from the very beginning right when it's being designed to you know when even after it's deployed so the need for diversity and bringing in these focus groups to even you know at any stage is going to help make the product better all right so I guess the the big question is the future of generative AI in terms of its trust level for organizations where do you sense that's headed it seems like it's a really complex question because I'm not sure we should ever reach 100% % Trust of generative AI we can we can believe in it we can use it we can Corral it and try to regulate it uh I don't know if I want us to 100% trust it that's my personal opinion of course but you know what's your I don't think we'll reach there uh you know in the near future look I I like to give this example right we're in this interesting time and I look at it as a huge opportunity for Our Generation right there are three parlel streams the first stream is you know think of it as the F the first car engine has been developed and and you know the car you know we are able to get from point A to point B faster right the car is still being developed but it's already on the road that's the second stream right where you so you're take you know AI research is happening but you're taking that and applying in the real world but when you take a still developing car in the real Road in the real world there are going to be you know there are no speed limits roads don't exist um you know horse carriage Riders uh you know the drivers their work is at risk right have to adapt the poor the poor Carriage Riders sure yeah there are no Lanes there is nothing so we have to build the roads and that's the third stream of regulations best practices you know making sure there are seat belts and speed limits and you know roads have built the infrastructure right so I think we're in this interesting phase James that we can actually influence and shape and set this up for the future but at the same time that for as long as AI is as a core technology is still growing evolving the nuances around it you know what kind of roads are needed that will keep changing so you know we are nowhere near close to reach the peak of AI research you know we you know and think about you know even if we reach that Peak you know what happens when Ai and Quantum come together right there are so many ways you know this technology is going to continue and evolve and so you know the speed limits the regulations will have to continue and evolve over the you know at least in the near term future but for me the survey reflects that you know people are you know have reached that point where they understand the tremendous value AI brings while also understanding that there are nuanced risks that come with it so that just that awareness and knowledge is the first step for us to start taking proactive steps to address it well all right and you I think you used the word regulation in there at one point which is interesting question it may be off topic but you know let's let's see where this goes in that um I think governments I know there was a I live here in California and there was a veto of an AI Bill recently of course it'll be Rewritten and it's just it's just one effort you know ongoing effort to regulate AI I think the government might have a hard time regulating AI partially because the companies that are building are well funded and very agile government sometimes can't move quite as quickly as those companies can so uh if if AI outruns regulation which may or may not happen we don't know um it's it's going to be a problem with the trust right so so fully regulating AI is going to be a particularly difficult issue agree disagree uh I think it'll take time the third stream that I go talked about is going to be the slowest because you want to put out regulations that's going to curb Innovation so think about it the first car came on the road in the early 1900s seat belts became mandatory much much later like 60 or 70 years later right because you want regulations that are thoughtful that don't curve Innovation but at the same time protect and keep the you know humans safe and secure right so I I I I do think it's going to be um it's going to you know take that balanced approach but at the same time I'm also seeing organization especially the ones who that are in highly regulated Industries uh you know getting proactive about self-regulating adapting existing regulations to address the implications of AI technology so we are seeing some best practices evolving but at the same time as long as at first stream you know where AI research is going full steam ahead there is going to be a catch up in both the application side and the regulation and R side all right so is there anything that particularly worries you or that you would you would share your worry with with managers in terms of generative Ai and Trust if you'd say oh please be careful of this someone might be worried about it is is is there a worry there in terms of generative a and Trust I think the concern is if you uh are you know if you if you don't go beyond finding the risk to actually solving for it thinking that you know oh you know and I think we past that phase but James if you remember a few years ago there was a lot of discussion around bias and and the need for fairness and somehow that has been replaced Now by by hallucination and safe and secure uh which I think is the right approach which in terms of being aware of all the risk understanding you know how to solve for it mitigate it based on what you know it's not a 100% solved problem right it is going the technolog is going to continue to evolve the risks are going to continue to evolve so there is it now more than ever managers and leaders need to be more agile need to build an ecosystem of Partners from whom they can learn and grow together I think that's the answer as far as there is a an answer I think that that's it uh bean I think you said it it was fascinating it's going to be a really interesting sector to follow and thank you so much for talking with us today uh please come back and talk with us again sometime thank you so much for having me James

This transcript was generated automatically from the video's captions and may contain errors.