Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Chatbot ethics, privacy and compliance

487 views

Published on

- Ákos Deliága - Talkabot -

Conversational user experience and mass personalization are hyped trends which are enabled by artificial intelligence and chatbots. Such tools speak with thousands of users simultaneously generating 10s of millions of conversations every month. How does an ethical chatbot behave? How's privacy and compliance handled? Ákos Deliága the managing director of Talk-A-Bot, a leading enterprise chatbot provider in CEE region will share practical experience and use cases.

IVSZ | EuDEco project
Data Economy Conference
Budapest, 2018. 01. 31.

Published in: Data & Analytics
  • Be the first to comment

Chatbot ethics, privacy and compliance

  1. 1. CHATBOT ETHICS, PRIVACY AND COMPLIANCE ÁKOS DELIÁGA – TALK-A-BOT CHRYS MARGARITIDIS - CEU
  2. 2. SOME EXAMPLES OF ETHICAL ISSUES: A. BIAS IN ALGORITHMS
  3. 3. BIAS IN ALGORITHMS • Guardian, December 7, 2016 • Who’s responsible? • How do you fix this?
  4. 4. METADATA
  5. 5. ARTIFICIAL INTELLIGENCE? • Where’s the ethical limit? Is there a limit?
  6. 6. BIG QUESTION
  7. 7. PREDICTIVE ANALYTICS: PREDICTIVE POLICING • The application of historical crime data to predict future criminal activity • Models depend mainly on past crimes (perpetrator, location, etc.) but can include seasonal information, salary schedules, etc. • Big Data innovation: application of artificial intelligence to large data sets • Goal: predict the location of future crimes
  8. 8. PREDICTIVE POLICING: CAVEATS • Assumption: the future is going to follow patterns in the past? • Oftentimes based on inaccurate sources • Not the most effective method to achieve the set goals • Costly to maintain • Becomes a self-fulfilling prophecy, resulting to the propagation of discriminatory practices
  9. 9. PREDICTIVE ANALYTICS: MASS SURVEILLANCE • Instead of using data to predict the future, use data to create a map of the present • Combines information from license plates, cameras, radiation sensors, and police databases • Big Data innovation: Provides the police real time access that can reveal information about a person’s location, activities and contacts • Goal: prevention of crime and effective, real-time dealing with escalating situations
  10. 10. MASS SURVEILLANCE: CAVEATS • Sacrifice important human rights for beneficial effects (increased safety and security)  Privacy  Freedom of expression  Freedom of conscience  Freedom of association • Slippery slope: from mass surveillance to predictive policing to pre-emptive punishing? • Humans are treated as data sets, without respect to their autonomy and identity
  11. 11. PRE-EMPTIVE SENTENCING? • Punish behavior that hasn’t occurred? • Causal elements in behavior • Presumption of Innocence • Fairness? • Can we maintain our identity if we cannot act out our choices? • Example: http://www.equivant.com/
  12. 12. 6 CRITICAL QUESTIONS FOR BIG DATA (BOYD & CRAWFORD) • Big Data changes the definition of knowledge • Claims to Objectivity and accuracy are misleading • Bigger data are not always better data • Taken out of context, Big Data loses its meaning • Just because it is accessible does not make it ethical • Limited Access to Big Data creates new digital divides
  13. 13. E N T E R P R I S E C H A T B O T P R O V I D E R TALK-A-BOT Creativity is our privilege, let the robots do the rest!
  14. 14. Traction 2016 EVENT: Web Summit Lisbon, Alpha track 2016 PRESS: Forbes online lead article Index.hu lead article 2017 EVENTS: Mobile World Congress Barcelona Axis Tel Aviv CeBIT Hannover Media Hungary IMPACT CEE Krakow Nestlé HackYourCompany Pitch competition first prize Techcrunch Disrupt San Francisco Visegrád in the Ring Reverse Pitch winner 2017 PRESS: Forbes HU: TOP 20 Fintech Startups to watch PCWorld: Top 10 Hungarian Startups Index.hu tech article Napi.hu interview 2018 PRESS: M2 Petőfi TV 2018 EVENTS: Bridge to MassChallange Warsaw We introduced chatbots to 2,000,000 users in just 16 months. We’re just getting started!
  15. 15. Europe’s first CHAT SHOP solution Europe’s first CHAT SHOP solution
  16. 16. • Unique chatbots • Scalable business model • Communication automation • Mass personalization • Chatbot marketing know-how • Native support for 100+ languages • In-house or private cloud installation • GDPR compliance Value Proposition Chatbot Anatomy
  17. 17. TAB FRAMEWORK BUSINESS LOGIC NLP AI DBDB CONNECTORSREPORTING INTEGRATION MONITORING TECHNOLOGY EXTERNAL USE CASES E-commerce Banking FAQ Hybrid chat Lead generation Self service Service messages Personalized content INTERNAL USE CASES IT helpdesk HR Recruitment Internal knowledge base Workflows MARKETING KNOW-HOW INTEGRATION Webshop CMS CRM API Webservice Database Workflows PaymentAuthentication Offering
  18. 18. BOT AND TALK-A-BOT: ETHICAL ISSUES • Who does the Bot serve, customer, intermediary or hiring company? • Who owns the data created? Is conversational information data? • Transparency in the processes above: everyone knows what belongs to who • Privacy: • can the Bot share the information with other Bots? Other humans? • Anonymity of customer • Right to be forgotten • User-confidentiality agreement • Commercial use of data? • Are humans allowed to abuse Bots? Vice versa? • Should Bots have a human face? Diversity • Bias in algorithms: discrimination
  19. 19. GOOD AND BAD BOTS • More internet traffic from bots than humans • Bot population: the majority is malicious • Good Bots • Commercial • Search engines • Feed fetchers • Monitoring
  20. 20. BOTS MISBEHAVING AND BEHIND BARS • https://twitter.com/tayandyou?lang=en • https://venturebeat.com/2016/09/05/this-is-the-first-chatbot-to-be-arrested/
  21. 21. Ákos Deliága Managing Director +36 70 294 0074 akos.deliaga@talkabot.net Creativity is our privilege LET THE ROBOTS DO THE REST!

×