WEBVTT 1 00:00:00.150 --> 00:00:04.350 In this lesson, we will learn about AI bots, 2 00:00:04.350 --> 00:00:07.380 artificial intelligence, or AI bots, 3 00:00:07.380 --> 00:00:09.960 our automated software programs 4 00:00:09.960 --> 00:00:12.690 that interact with users or systems 5 00:00:12.690 --> 00:00:15.420 to perform specific tasks. 6 00:00:15.420 --> 00:00:17.700 They may also serve as assistants 7 00:00:17.700 --> 00:00:19.560 and digital workers. 8 00:00:19.560 --> 00:00:23.790 AI bot concepts include access and permissions, 9 00:00:23.790 --> 00:00:27.990 guardrails, data loss prevention, or DLP, 10 00:00:27.990 --> 00:00:31.110 and disclosure of AI usage. 11 00:00:31.110 --> 00:00:35.790 Access and permissions refer to the controls put in place 12 00:00:35.790 --> 00:00:39.270 to ensure that AI bots only interact 13 00:00:39.270 --> 00:00:41.070 with data and systems 14 00:00:41.070 --> 00:00:44.790 that they are explicitly authorized to access. 15 00:00:44.790 --> 00:00:47.910 By doing this, access and permissions 16 00:00:47.910 --> 00:00:51.300 prevent unauthorized data manipulation 17 00:00:51.300 --> 00:00:53.460 and security breaches. 18 00:00:53.460 --> 00:00:57.750 Next, guardrails are predefined boundaries or rules 19 00:00:57.750 --> 00:01:01.230 that restrict the actions of AI bots. 20 00:01:01.230 --> 00:01:05.130 Then data loss prevention or DLP strategies 21 00:01:05.130 --> 00:01:07.530 are used with AI bots 22 00:01:07.530 --> 00:01:09.870 to prevent sensitive information 23 00:01:09.870 --> 00:01:13.230 from being accessed or leaked inadvertently. 24 00:01:13.230 --> 00:01:16.200 Finally, disclosure of AI usage 25 00:01:16.200 --> 00:01:18.060 involves informing users 26 00:01:18.060 --> 00:01:21.030 that they are interacting with an AI bot 27 00:01:21.030 --> 00:01:22.860 rather than a human. 28 00:01:22.860 --> 00:01:25.800 Let's learn more about access permissions, 29 00:01:25.800 --> 00:01:28.620 guardrails, data loss prevention, 30 00:01:28.620 --> 00:01:31.920 and the disclosure of AI usage. 31 00:01:31.920 --> 00:01:35.250 First, we have access and permissions. 32 00:01:35.250 --> 00:01:39.240 Access and permissions control what an AI bot can see 33 00:01:39.240 --> 00:01:41.400 or do within a system. 34 00:01:41.400 --> 00:01:45.420 So access requires setting explicit limits 35 00:01:45.420 --> 00:01:49.680 to make sure bots access only the data and systems 36 00:01:49.680 --> 00:01:52.590 they need to complete their tasks. 37 00:01:52.590 --> 00:01:56.040 Permissions then act as a safety barrier, 38 00:01:56.040 --> 00:01:58.980 ensuring AI bots can't overreach 39 00:01:58.980 --> 00:02:01.020 into sensitive areas. 40 00:02:01.020 --> 00:02:03.990 For example, in a healthcare setting, 41 00:02:03.990 --> 00:02:05.970 a patient assistance bot 42 00:02:05.970 --> 00:02:09.780 may only need access to general medical information 43 00:02:09.780 --> 00:02:11.910 to answer common questions, 44 00:02:11.910 --> 00:02:14.580 not specific patient records. 45 00:02:14.580 --> 00:02:17.820 So by carefully setting these permissions, 46 00:02:17.820 --> 00:02:20.430 the bot stays within its lane, 47 00:02:20.430 --> 00:02:22.590 interacting with data safely 48 00:02:22.590 --> 00:02:25.380 without risking unauthorized access 49 00:02:25.380 --> 00:02:27.180 to patient details. 50 00:02:27.180 --> 00:02:29.340 In industries where AI bots 51 00:02:29.340 --> 00:02:31.980 work closely with financial data, 52 00:02:31.980 --> 00:02:34.560 such as in the banking industry, 53 00:02:34.560 --> 00:02:37.950 access controls can be even more important. 54 00:02:37.950 --> 00:02:41.250 For example, an AI bot helping customers 55 00:02:41.250 --> 00:02:42.930 with account information 56 00:02:42.930 --> 00:02:45.300 would need tightly controlled access 57 00:02:45.300 --> 00:02:48.810 to prevent it from making unauthorized transactions 58 00:02:48.810 --> 00:02:51.390 or accessing sensitive financial records 59 00:02:51.390 --> 00:02:53.040 without permission, 60 00:02:53.040 --> 00:02:55.680 minimizing the risk of security breaches 61 00:02:55.680 --> 00:02:57.600 or data manipulation. 62 00:02:57.600 --> 00:03:00.750 So to prevent unauthorized access 63 00:03:00.750 --> 00:03:03.630 or misuse of sensitive information, 64 00:03:03.630 --> 00:03:06.300 access and permissions keep AI bots 65 00:03:06.300 --> 00:03:09.360 securely within their defined roles, 66 00:03:09.360 --> 00:03:13.590 protecting both data integrity and user privacy. 67 00:03:13.590 --> 00:03:16.200 Second, we have guardrails. 68 00:03:16.200 --> 00:03:20.520 Guardrails are boundaries set up around AI bots 69 00:03:20.520 --> 00:03:24.270 to keep them operating within intended limits. 70 00:03:24.270 --> 00:03:27.450 These rules help guide the bot's actions, 71 00:03:27.450 --> 00:03:31.380 ensuring they don't veer into unintended territory 72 00:03:31.380 --> 00:03:35.190 or make autonomous decisions that could be harmful. 73 00:03:35.190 --> 00:03:39.150 For example, in an online shopping application, 74 00:03:39.150 --> 00:03:41.310 a product recommendation bot 75 00:03:41.310 --> 00:03:42.690 might have guardrails 76 00:03:42.690 --> 00:03:45.600 that prevent it from suggesting unavailable 77 00:03:45.600 --> 00:03:47.700 or out-of-stock products. 78 00:03:47.700 --> 00:03:50.910 These guardrails help the bot provide accurate 79 00:03:50.910 --> 00:03:54.330 and reliable information to the users, 80 00:03:54.330 --> 00:03:58.410 preventing misunderstandings or dissatisfaction. 81 00:03:58.410 --> 00:04:00.240 In a business context, 82 00:04:00.240 --> 00:04:01.950 consider a cleaning bot 83 00:04:01.950 --> 00:04:04.140 in a Sam's Superstore, 84 00:04:04.140 --> 00:04:07.260 driving around and cleaning the floors. 85 00:04:07.260 --> 00:04:09.900 Guardrails are set to prevent this bot 86 00:04:09.900 --> 00:04:13.080 from colliding with customers or merchandise, 87 00:04:13.080 --> 00:04:16.890 ensuring it navigates safely around the store. 88 00:04:16.890 --> 00:04:19.890 These guardrails define specific paths 89 00:04:19.890 --> 00:04:21.750 or areas for cleaning, 90 00:04:21.750 --> 00:04:24.660 restrict its movements around crowded aisles, 91 00:04:24.660 --> 00:04:26.970 and ensure it stops or reroutes 92 00:04:26.970 --> 00:04:29.160 when encountering obstacles. 93 00:04:29.160 --> 00:04:31.320 By establishing these boundaries, 94 00:04:31.320 --> 00:04:34.260 the store guarantees that the bot performs 95 00:04:34.260 --> 00:04:36.450 its cleaning duties efficiently 96 00:04:36.450 --> 00:04:39.450 without compromising customer safety 97 00:04:39.450 --> 00:04:42.150 or risking damage to merchandise. 98 00:04:42.150 --> 00:04:45.780 So by having predefined rules in place, 99 00:04:45.780 --> 00:04:48.870 companies ensure their bots remain helpful 100 00:04:48.870 --> 00:04:50.850 and don't step into areas 101 00:04:50.850 --> 00:04:53.040 they are not programmed to handle, 102 00:04:53.040 --> 00:04:57.180 maintaining quality and safety in their interactions. 103 00:04:57.180 --> 00:05:01.530 Third, we have data loss prevention, or DLP. 104 00:05:01.530 --> 00:05:05.400 Data loss prevention measures are deployed with AI bots 105 00:05:05.400 --> 00:05:07.770 that handle sensitive information 106 00:05:07.770 --> 00:05:11.940 to prevent unintended exposure of sensitive data. 107 00:05:11.940 --> 00:05:13.440 DLP strategies, 108 00:05:13.440 --> 00:05:16.920 like encryption or restricted access policies, 109 00:05:16.920 --> 00:05:19.470 ensure that sensitive information 110 00:05:19.470 --> 00:05:22.050 stays within the authorized limits 111 00:05:22.050 --> 00:05:25.710 and is protected from unauthorized exposure. 112 00:05:25.710 --> 00:05:28.470 For instance, a banking chatbot 113 00:05:28.470 --> 00:05:31.050 might be designed with DLP measures 114 00:05:31.050 --> 00:05:32.760 to restrict any sharing 115 00:05:32.760 --> 00:05:35.190 of a user's financial details 116 00:05:35.190 --> 00:05:37.620 without proper verification. 117 00:05:37.620 --> 00:05:40.590 This helps in safeguarding user data 118 00:05:40.590 --> 00:05:43.470 and building trust in the bot's usage 119 00:05:43.470 --> 00:05:45.840 with insensitive environments. 120 00:05:45.840 --> 00:05:47.490 In a corporate environment, 121 00:05:47.490 --> 00:05:50.520 an HR department might use an AI bot 122 00:05:50.520 --> 00:05:53.460 to assist employees with common inquiries 123 00:05:53.460 --> 00:05:55.380 about company policies, 124 00:05:55.380 --> 00:05:58.320 benefits, or leave balances. 125 00:05:58.320 --> 00:06:00.690 DLP measures would be employed 126 00:06:00.690 --> 00:06:03.450 to ensure the bot cannot access 127 00:06:03.450 --> 00:06:05.730 or share personal information, 128 00:06:05.730 --> 00:06:07.680 like Social Security numbers, 129 00:06:07.680 --> 00:06:11.130 salary details, or medical leave records, 130 00:06:11.130 --> 00:06:15.780 unless the employee has gone through identity verification. 131 00:06:15.780 --> 00:06:19.470 This helps prevent sensitive human resources data 132 00:06:19.470 --> 00:06:22.110 from being disclosed inadvertently 133 00:06:22.110 --> 00:06:24.480 or to unauthorized users. 134 00:06:24.480 --> 00:06:27.060 So to prevent accidental leaks 135 00:06:27.060 --> 00:06:30.750 or unauthorized access to sensitive information, 136 00:06:30.750 --> 00:06:34.620 DLP strategies ensure that AI bots 137 00:06:34.620 --> 00:06:37.020 operate within strict boundaries, 138 00:06:37.020 --> 00:06:40.680 keeping data secure and interaction safe. 139 00:06:40.680 --> 00:06:45.480 Fourth and last, we have disclosure of AI usage. 140 00:06:45.480 --> 00:06:49.320 Disclosure of AI usage involves letting users know 141 00:06:49.320 --> 00:06:52.080 they are interacting with an AI bot 142 00:06:52.080 --> 00:06:53.880 rather than a human. 143 00:06:53.880 --> 00:06:57.600 This is essential for transparency and trust. 144 00:06:57.600 --> 00:06:59.550 Disclosure of AI usage 145 00:06:59.550 --> 00:07:01.800 allows users to understand 146 00:07:01.800 --> 00:07:04.380 the nature of their interaction 147 00:07:04.380 --> 00:07:07.290 and set appropriate expectations. 148 00:07:07.290 --> 00:07:10.470 For instance, a customer service chatbot 149 00:07:10.470 --> 00:07:12.390 on a retail website 150 00:07:12.390 --> 00:07:15.630 could display a message at the start of a conversation 151 00:07:15.630 --> 00:07:18.480 stating, "I am an AI assistant, 152 00:07:18.480 --> 00:07:21.420 here to help you with product inquiries." 153 00:07:21.420 --> 00:07:24.750 Such clear disclosure builds transparency 154 00:07:24.750 --> 00:07:26.610 and makes users more comfortable 155 00:07:26.610 --> 00:07:28.530 interacting with the bot. 156 00:07:28.530 --> 00:07:30.090 In healthcare settings, 157 00:07:30.090 --> 00:07:34.020 disclosure is especially important when using AI bots 158 00:07:34.020 --> 00:07:35.880 for patient interactions. 159 00:07:35.880 --> 00:07:38.160 For instance, if a healthcare bot 160 00:07:38.160 --> 00:07:40.290 provides a general wellness advice 161 00:07:40.290 --> 00:07:43.140 or answers to common medical questions, 162 00:07:43.140 --> 00:07:45.450 patients should be clearly informed 163 00:07:45.450 --> 00:07:47.850 that they are speaking to an AI, 164 00:07:47.850 --> 00:07:50.670 not a licensed medical professional. 165 00:07:50.670 --> 00:07:53.460 This helps set accurate expectations 166 00:07:53.460 --> 00:07:56.550 as the bot may lack the depth of expertise 167 00:07:56.550 --> 00:07:58.080 or personal insight 168 00:07:58.080 --> 00:08:00.060 that a doctor would provide. 169 00:08:00.060 --> 00:08:02.820 So by disclosing AI usage, 170 00:08:02.820 --> 00:08:06.360 businesses help users approach the bot's responses 171 00:08:06.360 --> 00:08:09.630 with the right level of caution and understanding, 172 00:08:09.630 --> 00:08:11.910 supporting trust and clarity 173 00:08:11.910 --> 00:08:14.160 in the service experience. 174 00:08:14.160 --> 00:08:19.160 So, remember, artificial intelligence or AI bots 175 00:08:19.257 --> 00:08:21.330 are automated programs 176 00:08:21.330 --> 00:08:24.480 designed to perform specific tasks, 177 00:08:24.480 --> 00:08:27.510 often simulating human interactions. 178 00:08:27.510 --> 00:08:30.420 They are commonly used as digital assistants 179 00:08:30.420 --> 00:08:33.030 or workers in various settings, 180 00:08:33.030 --> 00:08:36.900 including customer service and data management. 181 00:08:36.900 --> 00:08:38.400 To work securely, 182 00:08:38.400 --> 00:08:42.210 AI bots are set up with access and permissions, 183 00:08:42.210 --> 00:08:44.640 which control what data and systems 184 00:08:44.640 --> 00:08:46.320 they can interact with, 185 00:08:46.320 --> 00:08:49.590 helping prevent unauthorized access. 186 00:08:49.590 --> 00:08:53.460 Next, guardrails define safe boundaries, 187 00:08:53.460 --> 00:08:57.450 guiding bots to stay within their designated tasks. 188 00:08:57.450 --> 00:09:00.420 Then data loss prevention strategies 189 00:09:00.420 --> 00:09:02.490 protect sensitive information 190 00:09:02.490 --> 00:09:04.800 from being accidentally accessed 191 00:09:04.800 --> 00:09:06.600 or leaked by the bot. 192 00:09:06.600 --> 00:09:10.320 And finally, disclosing AI usage to users 193 00:09:10.320 --> 00:09:13.650 promotes transparency and builds trust 194 00:09:13.650 --> 00:09:16.950 by letting them know they're interacting with a bot 195 00:09:16.950 --> 00:09:18.573 rather than a human.