Intelligence and creativity: a winning partnership?

Q1) Providing the tools to be more creative, for example, word processors created a whole new group of writers, #AI has the ability to provide new tools - #saschat

A1: It can free humans up from the mundane, from the calculations. It can identify an incredible number patterns of potential interest for us to further examine. It can show where a certain process leads to in the extreme. #SASCHAT

A1 I think rather the opposite when AI are supporting automation of tedious work for us - and we can devote ourselves into better things being creative #AI #SASchat

A1: #AI should take human creativity to another level. It would probably redefine creativity, whatever is possible through AI is no more human creativity... #SASchat

A1 As we've seen w/ Go, #AI systems can show us new, unorthodox, "creative" ways of solving a problem. https://www.theverge.com/2017/10/18/16495548/deepmind-ai-go-alphago-zero-self-taught … #SASchat

A1: I would have to turn that around. Human creativity impacts AI. Humans evolve over millions of years! The creativity is already there. We refine AI in more interesting ways as compute power grows. #saschat

A1: An #AI will never be Chief Innovation officer, but “a promising process for innovative problem solving is a human computer collaboration where each partner assists the other ...”. Very interesting paper on the subject: http://bit.ly/2HBIDDz  #saschat

I definitely want to echo @leosadovy. The automation enabled by AI potentially frees us up to focus on more visionary/creative activities. #SASchat

RT @thoughts_Ash: A1: #AI should take human creativity to another level. It would probably redefine creativity, whatever is possible through AI is no more human creativity... #SASchat

RT @jaredpeterson: A1 As we've seen w/ Go, #AI systems can show us new, unorthodox, "creative" ways of solving a problem. https://www.theverge.com/2017/10/18/16495548/deepmind-ai-go-alphago-zero-self-taught … #SASchat

RT @rosenjosefin: A1: An #AI will never be Chief Innovation officer, but “a promising process for innovative problem solving is a human computer collaboration where each partner assists the other ...”. Very interesting paper on the subject: http://bit.ly/2HBIDDz  #saschat

A1: Yes - agree with @GrayColinAI and @leosadovy - Freeing us up from the mundane provides room for our creative abilities to flourish. This partly assumes, though, that we will have authority to point these AI capabilities at different challenges dynamically. #saschat

Q1: I think #AI is impacting human creativity by finding ways to address not only the complex, but also the mundane. If we can discover ways to #automate tedious things, we can be more #creative. Tedium and boredom are detrimental to creativity and #innovation. #SASChat

A1: Yes, #AI certainly frees up time for us to use our human abilities like for instance #creativity & #empathy, where an AI many times struggle. Also since AI can give you a lot of insights that will most likely give you new ideas and trigger your creativity. #saschat

A1: It can lead your thinking a step further with suggestions about songs you might like or ways to enhance a photo. Your own creativity can then start from a stage ahead and move further. #SASchat

AI shows impressive progress but I have yet to see a truly creative behavior from an AI in the sense that no rules are applied #saschat

A1: and what used to be AI we don't consider AI anymore, like playing chess better than the best human players #AI #saschat

A1: AI apps have subtle effects on human creativity, already pervasive: for a friend's birthday, some online shopping for a gift, a custom card made with https://deepdreamgenerator.com/  -- in both cases AI augments creative responses #SASchat

I agree, #AI is (usually) more enrichment than danger. #saschathttps://twitter.com/jaredpeterson/status/987317212032663552 …

A2 not sure public will trust #AI, but rather the organisations demonstrating responsible use of it. #saschat

A2a: Maybe two kinds of trust - “Trust this bridge won’t fall down”, and “Trust you”. Trust of the bridge type will likely come with experience, ubiquity and time. Trust of the second type is trusting motivations. This is going to be harder to achieve. #SASCHAT

True - the winning concept is definately when man and machine work together! #saschat #AI

A2b: Trusting motivations could be achieved by demonstrating the AI has no motivations. Or in cases where it acts as a middle-man trustee, like a lawyer or bank. But still, what about the corp/org behind it? You’ll trust the AI only if you already trust the organization. #SASCHAT

A1: it is like compare movies from 1960 to ones in 1990... think of Jurassic Park... earlier, imagine what can be achieved, now anything can be achieved, let us see what you can imagine... #saschat

A1: as of a metaphor to getting rid of mundane tasks, like solving a Rubik's cube in .6 seconds https://www.youtube.com/watch?v=iBE46R-fD6M … #saschat

A2 As with a lot of things in life, education is the key. The public needs to know what the true possibilities of AI are. Ignorance breeds fear. #SASchat

A2: It takes Fair, Accountable and Interpretable/Transparent#AI. That is key for the public to fully trust #AI... #FATML #XAI #saschat

A1: #AI do not deal with human creativity at all. #AI can just automate certain human tasks but rather a few today like image recognition, decisioning based on classification, text analysis #saschat

a2) I guess the public should be confirmed that AI is not about to remove their chance of making a decent life for themselves #AI will remove a lot of jobs #saschat #AI

A1: looking at industry, in health care AI is extending the ability of doctors to perceive, e.g., for radiology. Getting inside that feedback loop for a professional's is another example of augmenting creativity #SASchat

A2 The public needs to see a steady drumbeat of good stories. Things that are working, things that are helping. We love to the negatives and often don't praise the positives. #SASchat

A2 - In one aspect, I think it will take some assurance that jobs won't be taken away by AI. So this will require widespread AI to enable work that appeals to everyone. #saschat

A2: Transparency about when a communication is #AI rather than human would help. Chatbots can feel weird. Is this really a person? People don't want to be put in a position to undertake the Turing test. #SASchat

A2: I believe that people will trust AI when they experience a well positioned AI solution. E.g. a support chat in a specific industry/company where AI actually has the ability and environment to succeed in the support function #saschat

A2 Teach, inform and let people become aware of what AI can do and how – we can prevent this by enlightning the public - we have an obligation to do so!! #AI #SASchat

#saschat but we'll know #AI has reached full maturity when it refuses to classify any more 'cute' kitten images! https://twitter.com/alexge1/status/987318581053489152 …

A2 As AI evolves, there will continue to be distrust of recent innovation. Then people get used to things that were previously "creepy"; it becomes part of everyday life. CallerID was creepy, once-upon-a-time. So were targeted ads. OK, those still are creepy. #saschat

A2 Efforts on model explainability (FATAI) should help as well. Understanding the black box may put our fears at ease? #SASchat

A2) #AI and ethics should not be strangers. The public should expect #AI to improve lives for ALL - not the many, and especially not the few #saschat

A2: “When accuracy outpaces interpretability, human trust suffers, affecting business adoption, regulatory oversight, & model documentation" says @jpatrickhall & @Navdeep_Gill_ in this really nice, free ebook with @OReillyMedia https://oreil.ly/2HgHmmd  #AI #saschat #XAI #FATML

#saschat Good point, it would be interesting to stand up a survey "Do you trust AI" next to the survey questions "Do you trust organizations X, Y, and Z"? #saschat

A2: Or will it primarily *change* a lot of jobs? As new technology always has done. Else nobody should have a job 200 years after the industrial revolution? #AI #saschat

A2: Trust is earned and not usually automatically conferred. Customer-centered practices that are ethically informed will drive standards for trusting #AI #SASchat Article explains more: https://www.sas.com/en_us/insights/articles/analytics/artificial-intelligence-ethics.html …https://twitter.com/SASsoftware/status/987318219311669248 …

A2: Let's tell them what it WILL do - not what it can do... there's the will of the few controlling the #AI engines to be truly ethical #saschat

A2: #AI will be trusted by the public if it is regulated and certified by independent organization based on interpretability and transparency #saschat

Q2: trust is slippery; e.g., we can ask for model transparency and interpretation, which in general is good -- although almost no two verticals have the same definition for `interpretation` #SASchat

A2: As long as AI is driven in a scientist and/or developer's universe, it is difficult to identify with, for regular people. AI tend to be too academic - when an AI engine is spiced up with some low tech business rules and a meaningful user interface it gains trust #saschat

A2 We have to be honest as to how we represent this technology to the public. These systems are often limited by the data they've seen. Biased decisions can easily creep in if we're not paying attention. #SASchat

A2. Governments need to establish #AI ethics rules such as recently proposed in UK https://www.parliament.uk/business/committees/committees-a-z/lords-select/ai-committee/news-parliament-2017/ai-report-published/ … #SASchat

A2 When #AI is used for the good and beneficial for of the most trust can be obtained easily on the contrary when the effect is partisan and the benefit accelerates unfairness then it creates suspicion and mistrust #SASchat

A1: Right, but it depends on the application, right? A weather app for example, do we care how it's done as long as it's accurate? Like "How do planes fly?" #AI #saschat

A2 - @leosadovy Interesting because motivations makes us human, and understanding motivations can lead to trust or distrust. Perhaps, then if we trust in the numbers, we can trust in AI. #saschat

A4 While I would argue we're still trying to figure out the exact applications that are transformational, I think that conversational assistants fueled by AI advancements in #NLP are awesome. #SASchat

A4 My close 2nd place (for selfish reason) favorite AI advancement is self-driving cars. (YOU DRIVE! I'm gonna take a NAP!) #saschat

Q5: Exploring Mars! #SASchat

#saschat how about being able to order a pizza or call a cab by just talking to a device the size of a coke can. https://twitter.com/alexge1/status/987323951574605826 …

A4: I think one of the biggest accomplishments of #AI to date is that it's become a ubiquitous topic in most households. Listening to my kids "train" #Alexa to better understand their voices so they can pose questions about weird topics and get good responses is cool. #SASChat

A5: I am very confident that AI one day will be able to identify people - like: Really identify people, not only by their biometrics but also by interviewing them and recognizing their thinking.. That would be a first at the Border Control #saschat

A4: This is a nice example: "Mind-reading #AI works out what music is playing in your head - this may restore communication for those paralyzed" http://bit.ly/2qMgxep  #saschat

A5 Many #AI opportunities in #Healthcare: improve patient care and wellbeing mining digital medical records, accelerate illness diagnosis (speeding up innovative technologies), new drugs discovery and development, strengthen medical imaging diagnosis #SASchat

A5: Geospatial image analysis. Better speech recognition and better language translation. More robust margin-of-safety and failure-mode analysis. Complex time-series decomposition and analysis. #SASCHAT

A4 #saschat Biggest AI accomplishment is the accuracy with which #DeepLearning can perform image classification, object detection, facial recognition and so forth. So many opportunities especially around #iot and security.

A5: I would say it is yet ‘unknown’... #Sasci #SASchat

A4: In my little world, the biggest #AI accomplishment to date is Siri apparently learning to understand how I pronounce my friend's name when I dictate a text to him while driving. #LittleVictories #SASchathttps://twitter.com/SASsoftware/status/987322245587587072 …

#saschat A3) AI is singing with just Notes and Lyrics as inputs #creativity #AI https://youtu.be/HANeLG0l2GA 

Now if these AIs encouraged family interaction... "Aladdin says get of those screens kids" - This gets back to inspiring trust in AI! #saschat

A3: we're approaching a point nicknamed "The Intertwingularity", and example use of @ProjectJupyter for LIGO, LSST, CERN, etc., shows what can happen for truly large scale scientific collab -- using these approaches, to leverage machine learning #SASchat

A5: We are producing tremendous amounts of unstructured data - it is estimated that 80 % of an organizations data is unstructured. A lot of that data is not used for #analytics. There is a great untapped potential using more and more of this data in #AI applications. #saschat

A5: I don't think the art and politics is an area where #AI could play a role because it is linked to human #social changes #saschat

A5 As someone who works for an analytics company, I'm really excited by the possibilities of conversational assistants tapping into analytics thus making answers accessible to more people. #SASchat

A5: Biggest achievement - I just found it... in an IKEA store.. https://www.theguardian.com/science/2018/apr/18/defeated-by-ikeas-flatpack-call-in-the-robots … #saschat

A5 this has just started - there are tons of new applications and opportunities out there - makes me happy just thinking about it - excited!! #AI #SASchat

A5: There is so much opportunity for #AI to be used in #disaster response and other #humanitarian areas. We can analyze the data faster, find patterns faster, and deploy #aid faster. We could save more lives and improve the lives of so many people. #data4good #SASChat

You are on a roll today, @leosadovy. And I bet we could get some pretty decent AI haikus going... computers would be great at putting 5-7-5 syllables together. Then people could help tune the results a bit. A perfect collaboration! #saschat

#saschat I believe they can go simultaneously. Scaring people from using those tools can delay things. As you said, this area is moving very fast. One reason is there are so many people are working actively in this area. It is very recent that people started to focus on #XAI.

Q4: biggest accomplishment to date for AI ? I'd point to automated machine translation of human language, e.g. Google Translate. Few things are more "creative" or complex than human language. Also, consider long-term outcomes for that work on software engineering ... #SASchat